City Research Online

City, University of London Institutional Repository

Citation: Li, F. (2015). Context-Aware Attribute-Based Techniques for Data Security and
Access Control in Mobile Cloud Environment. (Unpublished Doctoral thesis, City University
London)

This is the accepted version of the paper.

This version of the publication may differ from the final published version.

Permanent repository link: https://openaccess.city.ac.uk/id/eprint/11891/

Link to published version:

Copyright: City Research Online aims to make research outputs of City,

University of London available to a wider audience. Copyright and Moral Rights
remain with the author(s) and/or copyright holders. URLs from City Research
Online may be freely distributed and linked to.

Reuse: Copies of full items can be used for personal research or study,
educational, or not-for-profit purposes without prior permission or charge.
Provided that the authors, title and full bibliographic details are credited, a
hyperlink and/or URL is given for the original metadata page and the content is
not changed in any way.
City Research Online: http://openaccess.city.ac.uk/ [email protected]
Context-Aware Attribute-Based Techniques for Data
Security and Access Control in Mobile Cloud
Environment

A Thesis Submitted to
City University London, School of Engineering and Mathematical Sciences
In Fulfillment of the Requirements for the Degree
Doctor of Philosophy in
Information Engineering

By

FEI LI

April, 2015
Table of Contents

List of Figures………………………………………………………………………………………………….....V
List of Tables……………………………………………………………………………………………………VIII
Acknowledgements…………………………………………………………………………………………...IX
Declaration…………………………………………………………………………………………………………X
Abstract……………………………………………………………………………………………………………..XI
Notation and Abbreviation……………………………………………………………………………….XII
Publications…………………………………………………………………………………………………..…XVI

1 Introduction.............................................................................................................................1
1.1 Motivation .................................................................................................................2
1.1.1 Problems with the Current Technologies ..............................................................4
1.2 Contributions of the Thesis .......................................................................................7
1.3 Outline of the Thesis .................................................................................................9
2 Identity, Privacy, and Security in Mobile Cloud Environment ...............................................12
2.1 Mobile Cloud Computing.........................................................................................13
2.1.1 Concept of Mobile Cloud Computing ..................................................................13
2.1.2 Advantages of Mobile Cloud Computing .............................................................19
2.1.3 Issues of Mobile Cloud Computing ......................................................................21
2.2 Case Study ...............................................................................................................24
2.2.1 Existing Data Protection Laws .............................................................................27
2.3 Security Concepts, Technologies and Mechanisms .................................................30
2.3.1 Security Technologies and Mechanisms ..............................................................31
2.3.2 Protocols and Standards ......................................................................................36
2.4 Summary..................................................................................................................39
3 Identity Management Systems..............................................................................................41

I
 3.1 Single-Sign-On (SSO) and Federation.......................................................................42
3.2 Microsoft .NET Passport: .........................................................................................43
3.3 The Liberty Alliance(Kantara) ..................................................................................46
3.4 OpenID.....................................................................................................................48
3.5 Higgins .....................................................................................................................51
3.6 OAuth ......................................................................................................................53
3.7 Comparison and Literature Review .........................................................................55
3.8 Conclusion ...............................................................................................................61
4 Access Control Technologies .................................................................................................62
4.1 Access Control Models ............................................................................................63
4.1.1 Discretionary Access Control ...............................................................................64
4.1.2 Mandatory Access Control ..................................................................................66
4.1.3 Role-Based Access Control ..................................................................................68
4.1.4 Attribute Based Access Control ...........................................................................71
4.2 Privacy-Preserving Languages .................................................................................74
4.2.1 The Platform for Privacy Preferences ..................................................................75
4.2.2 Enterprise Privacy Authorization Language .........................................................76
4.2.3 Extensible Access Control Markup Language ......................................................76
4.3 Attribute Based Encryption .....................................................................................83
4.3.1 Key-Policy Attribute-Based Encryption ................................................................86
4.3.2 Ciphertext-Policy Attribute-Based Encryption.....................................................87
4.3.3 Multi-Authority Attribute-Based Encryption .......................................................89
4.3.4 Challenges ...........................................................................................................95
4.4 Conclusion ...............................................................................................................97
5 User-Centric Attribute-Based Access Control Model Using XACML ......................................99
5.1 Architecture of Policy-Based User-Centric Approach ............................................100
5.1.1 System Initialization...........................................................................................104
5.1.2 Design of the Model ..........................................................................................105
5.2 Policy Evaluation Component (PEC) ......................................................................110
5.3 Security Evaluation ................................................................................................115
5.3.1 Protocols on Authentication ..............................................................................115

II
 5.3.2 Security Analysis ................................................................................................117
5.3.3 User-Centric Approach ......................................................................................122
5.3.4 Use Case Study ..................................................................................................123
5.4 Proof of Concept....................................................................................................125
5.4.1 Protocol Verification ..........................................................................................125
5.4.2 Implementation and Tests .................................................................................130
5.4.3 Sample Screenshots of the Client Application...................................................131
5.4.4 Possible Extension .............................................................................................136
5.4.5 XACML Message Standard .................................................................................136
5.5 Discussion ..............................................................................................................137
5.6 Conclusion .............................................................................................................139
6 Context-Aware Attribute-Based Encryption Schemes .........................................................140
6.1 Introduction ...........................................................................................................140
6.2 Context-Aware Single Authority Attribute-Based Encryption Scheme ..................143
6.2.1 Preliminaries......................................................................................................144
6.2.2 Construction ......................................................................................................146
6.3 Context-Aware Multi-Authority Attribute-Based Encryption Scheme ..................150
6.3.1 Preliminaries......................................................................................................151
6.3.2 Construction ......................................................................................................154
6.3.3 Security Analysis ................................................................................................159
6.3.4 Performance Analysis ........................................................................................165
6.3.5 Computational Complexity Analysis ..................................................................165
6.3.6 Communication Complexity Analysis ................................................................171
6.4 Low-Complexity Multi-Authority Attribute-Based Encryption Scheme.................172
6.4.1 Constructions ....................................................................................................173
6.4.2 Security Analysis ................................................................................................178
6.4.3 Performance Analysis ........................................................................................179
6.5 Conclusion .............................................................................................................183
7 Conclusions and Future Work .............................................................................................185
7.1 Summary and Conclusions ....................................................................................185
7.2 Recommendations for Future Work ......................................................................191

III
Bibligraphy:.............................................................................................................................193
Appendix ................................................................................................................................209
XACML Language ................................................................................................................209
XACML Policy ..................................................................................................................209
XACML Request ..............................................................................................................212
XACML Response ............................................................................................................213

IV
List of Figures

Figure 2.1Authorization Information of Instagram at a Facebook Account. ....................25

Figure 2.2Symmetric En/Decryption Process ...................................................................33

Figure 2.3Asymmetric En/Decryption process .................................................................34

Figure 3.1.NET Passport Authentication Process .............................................................44

Figure 3.2 The Abstract View of Liberty Architecture ......................................................46

Figure 3.3The Liberty Notation of Trust ...........................................................................48

Figure 3.4OpenID Authentication Process .......................................................................49

Figure 3.5Architecture of Higgins .....................................................................................52

Figure 3.6Workflow of OAuth Protocol ............................................................................54

Figure 4.1Illustration of the Elements in a Standards RBAC Model .................................69

Figure 4.2Illustration of a Role Hierarchy .........................................................................70

Figure 4.3 Data Flow of a Standard XACML Framework...................................................78

Figure 4.4Illustration of Attribute-Based Encryption Scheme ..........................................84

Figure 4.5Key-Policy Attribute-Based Encryption Scheme ...............................................87

Figure 4.6Ciphertext-Policy Attribute-Based Encryption Scheme ....................................88

Figure 4.7Illustration of Multi-Authority Attribute-Based Encryption .............................92

Figure 5.1 Framework of the proposed User-Centric Policy-Based Access Control Model

V
 Using XACML ..........................................................................................................102

Figure 5.2Illustration of the Message Flow of Proposed User-Centric Policy-Based Access

Control Model ........................................................................................................107

Figure 5.3 Architecture of the PEC .................................................................................111

Figure 5.4Illustration of Work Flow of Policy Evaluation ................................................113

Figure 5.5Authentication between User and SP ............................................................115

Figure 5.6Authenticaion for SP .......................................................................................116

Figure 5.7Authentication for User ..................................................................................116

Figure 5.8 Validation of the Proposed Model by Scyther Toll ........................................129

Figure 5.9Proof of Concept ............................................................................................130

Figure 5.10Login Page of the Client Application.............................................................132

Figure 5.11Requesting Page of the Client Application ...................................................132

Figure 5.12Information Gathering at PEC ......................................................................133

Figure 5.13Authorization Result of PEC Service .............................................................133

Figure 5.14Service Test Page ..........................................................................................134

Figure 5.15PEC Service Test ............................................................................................135

Figure 6.1The Framework of Single Authority ABE scheme. ..........................................144

Figure 6.2Anonymous Key Issuing Protocol ...................................................................154

Figure 6.3 Comparison of Computational Costs for Encryption Between Conventional

MA-ABE Scheme and the proposed Context-Aware MA-ABE Scheme ..................168

VI
Figure 6.4 Comparison of Computational Costs for Decryption Between Conventional

MA-ABE Scheme and the Proposed MA-ABE Scheme ...........................................169

Figure 6.5 Illustration of the proposed Framework of Low-Complexity Context-Aware

Multi-Authority ABE Scheme for Mobile Cloud Environment ................................173

Figure 6.6 Comparisons of Communications between Chase and Chow’s MA-ABE

Scheme and proposed Low Complexity MA-ABE Scheme .....................................181

VII
List of Tables

Table 3.1 Comparison of the Different Identity Management

3.1 55
Schemes

Comparison of Computational Cost of the Single Authority

6.1 ABE scheme and the Context-Aware Single Authority ABE 166
Scheme

Comparison of Computational Complexity of the proposed

6.2 Context-Aware MA-ABE Scheme and Chase and Chow‟s 167
MA-ABE Scheme

6.3 Time Complexity Measures for Tow Different Test Beds 168

Comparison of Number of Required Computations between

6.4 Chase and Chow‟s Conventional MA-ABE Scheme and the 182
Proposed Low Complexity MA-ABE Scheme

VIII
Acknowledgements

This thesis arose after years of research that has been done since I came
to City University London. First and foremost, I would like to express my
gratitude to my supervisor Professor Muttukrishnan Rajarajan, for his supervision,
advice and guidance from the starting of my research. With his extraordinary
experience and admirable insights, he has taught me about science, life, and
encouragement. It is my pleasure to work with you.
I also would like to acknowledge the research collaboration with Dr.
Yogachandran Rahulamathavan. Thanks for the invaluable advice and patience. I
am grateful to my colleagues and everyone for their support and help on my
research work.
By the completion of this thesis, I am approaching the end of a 20 years
long life for officially enrolled as a student. Looking back at these years I will
remember with affection the people I have met and the moments with joy and
happy along the way.
I owe my final thanks to my beloved parents, with their unconditional
support, love, and trust. This thesis is dedicated to them.

IX
Declaration

No portion of the work referred to in this thesis has been submitted in support
of an application for another degree or qualification of this or any other university
or other institute of learning. I hereby grant powers of discretion to the University
Librarian to allow this thesis to be copied in whole or in part without further
reference to the author. This permission covers only single copies made for study
purposes, subject to normal conditions of acknowledgement.

X
Abstract

The explosive growth of mobile applications and Cloud computing has enabled
smart mobile devices to host various Cloud-based services such as Google apps,
Instagram, and Facebook. Recent developments in smart devices‟ hardware and
software provide seamless interaction between the users and devices. As a result,
in contrast to the traditional user, the mobile user in mobile Cloud environment
generates a large volume of data which can be easily collected by mobile Cloud
service providers. However, the users do not know the exact physical location of
their personal data. Hence, the users cannot control over their data once it is stored
in the Cloud. This thesis investigates security and privacy issues in such mobile
Cloud environments and presents new user-centric access control techniques
tailored for the mobile Cloud environments.
Most of the work to date has tried to address the data security issues on the Cloud
server and only little attention has been given to protect the users‟ data privacy.
One way to address the privacy issues is to deploy access control technique such
as Extensible Access Control Markup Language (XACML) to control data access
on users‟ data. XACML defines a standard of access control policies, rule
obligations and conditions in data access control. XACML utilizes Extensible
Markup Language (XML) schema to define attributes of data requesters, resources,
and environment in order to evaluate access requests. A user-centric
attribute-based access control model using XACML which enables users to define
privacy access policies over the personal data based on their preferences is
presented.
In order to integrate the data security and user‟s privacy in mobile Cloud
environment, the thesis investigates attribute-based encryption (ABE) scheme.
ABE scheme enables data owners to enforce access policies during the encryption.
Context-related attributes such as requester‟s location and behavior are
incorporated within ABE scheme to provide data security and user privacy. This
will enable the mobile data owners to dynamically control the access to their data
at runtime. In order to improve the performance, a solution that offloads the
high-cost computational work and communications from the mobile device to the
Cloud is proposed. Anonymisation techniques are applied in the key issuing
protocol so that the users‟ identities are protected from being tracked by the
service providers during transactions. The proposed schemes are secure from
known attacks and hence suitable for mobile Cloud environment. Security of the
proposed schemes is formally analyzed using standard methods.

XI
Notations and Abbreviations

2PC Two-Phase Commit Protocol

3G Third Generation

4G Forth Generation

AES Advanced Encryption Standard

APPEL A P3P Preference Exchange Language

CH Context Handler

CIA Confidentiality Integrity Availability

CP-ABE Ciphertext-Policy Attribute-Based Encryption

DAC Discretionary Access Control

DES Data Encryption Scheme

DVLA Driver and Vehicle Licensing Agency

EC2 Elastic Compute Cloud

EPAL Enterprise Privacy Authorization Language

EPIC Electronic Privacy Information Centre

GID Global Identity

GPS Global Positioning System

HIPAA Health Insurance Probability and Accountability Act

HTML Hyper Text Markup Language

HTTP Hypertext Transfer Protocol

ID Identity

IdP Identity Provider

ID-FF Liberty Identity Federation Framework

XII
ID-WSF Liberty Identity Web Services Framework

ID-SIS Liberty Identity Services Interfaces Specification

IMEI International Mobile Station Equipment Identity

J2EE Jave 2 Platform Enterprise Edition

JAX-WS Java API for Web Services

JSON JavaScript Object Notation

JSP JaveServer Pages

KP-ABE Key-Policy Attribute-Based Encryption

LTE Long-Term Evolution

MA-ABE Multi-Authority Attribute-Based Encryption

MAC Media Access Control

ManAC Mandatory Access Control

NHS National Health Service

NI National Insurance

NIST National Institute of Standards and Technology

Organization for the Advancement of Structured
OASIS
Information Standard
OECD Organization for Economic Co-operation and Development

OMB U.S. Office of Management and Budget

P3P Platform for Privacy Preferences

PAP Policy Administration Point

PDP Policy Decision Point

PDS Personal Data Service

PEP Policy Enforcement Point

PEC Policy Evaluation Component

PHI Personal Health Information

XIII
PII Personally identifiable Information

PIP Policy Information Point

QoE Quality of Experience

QoS Quality of Service

RBAC Role-Based Access Control

RG Request Generator

RSA Rivest-Shamir-Adleman

SAML Security Assertion Markup Language

SHA Secure Hash Algorithm

SOA Service-Oriented Architecture

SOAP Simple Object Access Protocol

SP Service Provider

SPId Service Provider Identity

SSL Secure Sockets Layer

SSN Social Security Number

SSO Single-Sign-On

SSS Secret-Sharing Schemes

STA Semi-Trusted Authority

TLS Transport Layer Security

UDDI Universal Description, Discovery and Integration

URL Uniform Resource Locator

VP Validating Point

W3C World Wide Web Consortium

WLAN Wireless Local Area Network

WSDL Web Services Description Language

XIV
XACML eXtensible Access Control Markup Language

XML Extensible Markup Language

XV
Publications

The results of the research described in this thesis have been published in
the following papers:

[1] Fei Li, DasunWeerasinghe, Dhiren Patel, Muttukrishnan Rajarajan, “An

User-Centric Attribute Based Access Control Model for Ubiquitous Environment”

Mobile Computing, Applications, and Services. Springer Berlin Heidelberg, 2012:

361-367.

[2] DasunWeerasinghe, FEI LI, Muttukrishnan Rajarajan, “Novel Framework for

Secure Mobile Banking”, In Proceedings of the 5th International Conference on

Security for Information Technology and Communications. 31 May – 1June 2012,

Bucharest, Romania

[3] Fei Li, Yogachandran Rahulamathavan, Muttukrishnan Rajarajan, RCW Phan

“Low Complexity Multi-authority Attribute Based Encryption Scheme for Mobile

Cloud Computing” Service Oriented System Engineering (SOSE). 2013 IEEE 7th

International Symposium on, pp.573,577, 25-28 March 2013, San Francisco Bay,

USA

XVI
[4]Fei Li, Yogachandran Rahulamathavan, Mauro Conti, Muttukrishnan Rajarajan,

“LSD-ABAC: Lightweight Static and Dynamic Attributes Based Access Control

Scheme for Secure Data Access Control in Mobile Environment”39th Annual

IEEE Conference on Local Computer Networks, 8-11 September 2014, Edmonton,

Canada

[5]Fei Li, Yogachandran Rahulamathavan, Mauro Conti, Muttukrishnan Rajarajan,

"Robust Access Control Framework for Unified Communications Network"

Special Issue on Security and Privacy in Unified Communications: Challenges and

Solutions (under review)

[6] Yogachandran Rahulamathavan, Suresh Veluru, Jinguang Han, Fei

Li,Muttukrishnan Rajarajan, and Rongxing Lu, "User Collusion Avoidance

Scheme for Privacy-Preserving Decentralized Key-Policy Attribute-Based

Encryption", IEEE Transactions on Information Forensics and Security(under

review)

XVII
 Chapter 1. Introduction

1 Introduction

Mobile handsets have changed human life to a great extend during the last couple of

decades, and it was one of the biggest inventions of the early 1980s.The recent

technological advancements in mobile devices and wireless technology enabled

users to migrate to portable computers such as smartphones and tablet computers

from traditional desktop computers; hence, users can access online services via

mobile handsets from anywhere and anytime. This transformation increased the

user interaction with computers to a great extent. Various new online technologies

and services are introduced to adopt the change in computer landscape. Mobile

Cloud computing is one of the emerging technologies where the traditional Cloud

computing is being transformed to improve the mobile Internet user experience.

Mobile Cloud Computing is defined as the application of Cloud computing in

combination with mobile devices over mobile networks [1]. The combination of

Cloud computing and mobile technology improves the user experience and

productivity, particularly in the working environment [2], e.g. location-based

services such as Google Map services via mobile devices provide convenient

navigation services to mobile Cloud users. All the necessary distance and route

optimizations are done within the Cloud. Mobile devices are not involved into such

1
 Chapter 1. Introduction

high-performance computations. However, this change creates challenges in terms

of user data security and privacy. Within this context, this thesis discusses the

security and privacy perspective of this trend and proposes novel algorithms to

mitigate the risk in identity management and privacy protection. In the following,

various components of mobile Cloud computing are briefly discussed followed by

contributions and structure of this thesis.

1.1 Motivation

The recent technology advancements in mobile communication networks and the

increasing penetration of smartphones are transforming the mobile Internet and

empowering the users with rich mobile experience. Today‟s users prefer to access

online data and services, such as using Dropbox for Cloud storage, and sharing

photos with friends using Instagram via mobile devices [3]. This trend has fuelled

the need to transform the traditional Cloud computing into mobile Cloud

computing that can help to address some of the data processing challenges.

However, the limitations of smartphones such as onboard computing power,

battery time, and storage capabilities of mobile devices hamper their ability to

support the increasing sophisticated application demanded by users [4]. Mobile

Cloud computing is an emerging Cloud service model that combines the Cloud

computing paradigm with traditional mobile networks framework. Researchers

2
 Chapter 1. Introduction

embrace new paradigm of mobile Cloud computing such as mobile storage as a new

way to extend the capabilities of mobile devices and mobile platforms, which has

the potential to impact the business environment and provide more convenience to

individual's daily life.

When users are accessing Cloud based services through their mobile devices, the

Cloud service providers may collect data from the users for service delivery. Most

of today‟s Cloud service providers work in a collaboration environment. Consider

the following scenario, an Instagram user can either share his photos with other

preferred Instagram users, or publish the content on his Facebook page. In order to

do so, he has to authorize Instagram with certain access rights on his Facebook

account. There are two security concerns in this scenario:

(1) Identity leakage: Users‟ Facebook identity data will be collected by

Instagram. Service providers collect user‟s attributes or identity

related data during the service period. In mobile Cloud computing

environment, the user‟s real identity can be disclosed without

permission by a collusion attack from a number of service providers,

which is considered as a new security threat from mobile Cloud

computing.

3
 Chapter 1. Introduction

(2) Data and Privacy Protection: Users use the Cloud for online data

storage. The data managed by Cloud service providers consists of

different types of information, such as identity information, location

data, and medical records. This information is closely associated

with user privacy; hence, the access privileges to access this

sensitive information should be strictly controlled. However,

existing security approaches are not suitable for protecting the users'

privacy in a mobile Cloud computing environment.

The above security and privacy issues are new challenges in mobile Cloud

computing environment. Hence new security solutions are required to seamlessly

secure the data and online services delivered over the mobile Cloud platform, as

well as users‟ privacy.

1.1.1 Problems with the Current Technologies

Existing security mechanisms of current information systems are becoming

increasingly inadequate for today‟s complex mobile Cloud environments. The

majority of current security implementations are adopted from the traditional

4
 Chapter 1. Introduction

approaches, which are not suitable for mobile Cloud environment [3, 5-7]. The

major shortcomings of the existing security solutions are summarised below:

Lack of control for users to protect their data: Service providers often hold and

process user data. If such data are disclosed or abused by unauthorized service

providers, the mobile Cloud users may lose their competitive advantage or even go

out of business. Most of the security solutions implemented by the Cloud providers

do not allow the data owners to define the necessary access control policies to

restrict access for the personal data. Hence, tailored access control techniques are

required in order to restrict data access in Cloud computing in order to protect

user‟s privacy.

Lack of protection mechanisms against user identity theft: In general, mobile

users obtain different services through various Cloud service providers, e.g. when

a Facebook user launches a third-party app in the Facebook App store, the app

providers requests the users' identities and data access consent (i.e. to read the

friendship details, location information or publish data on the personal home page

etc.) in order to deliver the requested service. Personal data such as friendships

and location information can be considered as privacy information. Hence, each

app provider knows the users' partial identity and privacy information from their

Facebook profile.

5
 Chapter 1. Introduction

Today most of the social media apps are accessed from smartphones. Hence it is

possible to obtain real-time contextual attributes (i.e. location, network

connectivity, app usage)which are available in mobile Cloud environments. If two

or more service providers corporate together to combine a user‟s information that

they hold, the user‟s sensitive privacy information may be illegally shared without

the user‟s consent. In order to protect the privacy of the individual, untraceable

identities are required in mobile Cloud computing. The identity management

solutions should satisfy the requirements that user‟s real identity should be

protected and be hidden from the service providers.

The centralized architecture is inadequate: In real life, a number of authorities

hold different types of attributes for an individual. Higher education institutions

have people‟s educational records. Hospitals store their patients‟ medical records.

These organizations can be considered as attribute authorities who maintain a set of

attributes for the user. However, most of the current implementations have only one

trusted authority, which acts as the administrator of the whole system, maintains

identity management and issues decryption keys. It will become a bottleneck of the

system if there are a large number of users. Storing all users‟ information at one

central repository is not an ideal solution too. If a Cloud service provider is attacked

then all sensitive information of the users will be compromised.

6
 Chapter 1. Introduction

This thesis focuses on addressing the problems stated above, and proposes robust

solutions to overcome the current security vulnerabilities in the mobile Cloud

computing environment.

1.2 Contributions of the Thesis

This thesis contributes to the development of a framework for identification,

authorization, and data access control in mobile Cloud computing environments.

The current security technologies in Cloud environment are examined and a

context-aware Multi-Authority Attribute-Based Encryption (MA-ABE) scheme to

address the above security issues in mobile Cloud environment is proposed. The

MA-ABE scheme enables data owners to define access policies during encryption

and is flexible and practical for mobile Cloud computing environment. A number of

attribute authorities can work together in a collaboration environment, and issue

secret keys to the data owner to secure his personal data that is stored in the Cloud.

By capturing real-time context-related attributes via users‟ mobile devices, a

security solution can be achieved for authentication and authorization.

In order to restrict access for users‟ online data, an Attribute-Based Access Control

(ABAC) Model using Extensible Access Control Markup Language (XACML)

for mobile Cloud environment is presented. XACML is a privacy access policy

language which performs access control based on attributes. Using XACML allows

7
 Chapter 1. Introduction

access rules to be defined in a policy-oriented fashion, with policies being easily

updated as rules change and is suitable for mobile Cloud environment.

The following are the major contributions of this thesis:

- Multi-Authority Attribute-Based Encryption in mobile Cloud

computing: To protect the individuals data and privacy in mobile

Cloud environments, the MA-ABE scheme is used to enable data

owners to define access control policy while storing encrypted data

in the Cloud. A data requestor can only decrypt the data based on

his/her attributes. In this scheme, a number of attribute authorities

maintain different sets of attributes and issue decryption keys for the

user respectively. Users obtain decryption keys from each authority

based on their attributes. The model has removed the trusted

authority which is responsible for monitoring communications and

issuing decryption keys in existing MA-ABE systems. Thus, it

improves the security and system efficiency.

- Incorporation of context-related attributes to conventional ABE

shceme: The available contextual information in mobile Cloud

environment is used by data owners during the encryption and

decryption process, which strengthens the security and privacy

8
 Chapter 1. Introduction

protection in a more flexible, user-centric, and fine-grained manner.

Light-weight cryptographic algorithms are also designed for mobile

Cloud computing environment. By utilizing the Cloud infrastructure,

the light-weight cryptographic algorithm offloads the high-cost

computation and communication overheads to the Cloud.

- Anonymisation key issuing protocol: Anonymisation techniques

are involved in the key issuing process. An anonymous algorithm is

used to mathematically hide users‟ real identity. Thus, identities are

protected from collusion attacks (see in section 1.1.1) in mobile

Cloud environments.

- XACML-based access control model: A user-centric access

control model for mobile Cloud environment is proposed to address

user‟s data privacy issues in section 1.1.1. The model is based on the

eXtensible Access Control Markup Language (XACML) and

enables users to define their own access control policies to secure

their data from unauthorized access.

1.3 Outline of the Thesis

The thesis consists of six Chapters in total. Chapter 2 gives an introduction and

related definitions for mobile Cloud computing, identity, privacy and security

related technologies. Then main technological building blocks for protecting

9
 Chapter 1. Introduction

identity and data security such as cryptographic schemes, security protocols are

discussed in this Chapter. The existing identity management systems such as .NET

Passport, The Liberty Alliance, OpenID, Higgins and OAuth are examined in detail

in Chapter 3. In addition to this the comparisons of the widely used identity

management systems are also given in Chapter 3.

Chapter 4 discusses the guidelines of privacy protection, and discusses the current

approaches in privacy protection, e.g. solutions for privacy, data access control, and

then proceeds by explaining the main building blocks required for data access

control. Several widely used access control models are discussed followed by a

brief discussion of privacy protection languages. The current schemes of

attribute-based encryption are also examined. Finally, the requirements and

challenges for privacy-preserving and data access control for mobile Cloud

environment are discussed.

Chapter 5 presents a new user-centric attribute-based access control model using

XACML for mobile Cloud environments. The proposed framework enables users

to define their own access control polices for their personal data. Real-time

attributes are used to secure the data transactions. An attribute authority stores and

manages all the attributes used in this study. The proof of concept of the proposed

model is given at the end of the chapter.

10
 Chapter 1. Introduction

Chapter 6 proposes three different attribute-based encryption schemes for mobile

Cloud environments: 1) context-aware single authority attribute-based encryption

scheme 2) context-aware multi-authority attribute-based encryption scheme and 3)

low-complexity multi-authority attribute-based encryption scheme. In the

context-aware single authority ABE scheme, any context-related attributes that can

be captured from user‟s mobile devices are used during the encryption. In addition

to this a context-aware multi-authority attribute-based encryption scheme is

proposed in order to address issues of the single authority scenario. Finally, a low

complexity multi-authority attribute-based encryption scheme is presented to

offload heavy computation and communication tasks from the mobile device to the

Cloud. The Cloud infrastructure is utilized to improve the efficiency of

multi-authority schemes. The algorithm is designed as a lightweight solution which

can secure the data confidentiality and preserve the user‟s privacy.

Chapter 7 concludes by summarizing the contribution of the thesis and highlights

the possible future research directions.

11
2 Identity, Privacy, and Security in Mobile Cloud
Environment

This chapter reviews the concept of mobile Cloud computing, identity, privacy and

security as well as outlines important protocols used in this thesis. In particular,

new issues rose in identity management, privacy and security technologies in

mobile Cloud computing environment are discussed.

This chapter is organized as follows: the Section 2.1 describes the concept of

mobile Cloud computing followed by a case study of identity, privacy and security

technologies in mobile Cloud environment in Section 2.2. Several security services

and mechanisms which are currently implemented in mobile Cloud environments

are outlined in Section 2.3. The conclusions are drawn in Section 2.4.

12
 Chapter 2. Identity, Privacy, and Security in Mobile Cloud Computing

2.1 Mobile Cloud Computing

2.1.1 Concept of Mobile Cloud Computing

Mobile devices now provide online data and services to users from anywhere at any

time. Most of the UK mobile operators have established 3G mobile networks and

are now moving forward to 4G LTE networks which can provide much better

performance compared to the 3G networks [8]. The term mobile Internet refers to

the combination of the mobile communication and Internet [9]. Hence, mobile

users can enjoy much better mobile data services via mobile Internet. The number

of smartphone subscribers within the UK is more than 50% of standard mobile

users in 2013 and it is predicted to reach up to 75% by 2016 [10]. Due to the

development of wireless communication technology and Web application

technology, mobile Internet provides endless space for mobile network and a

platform for developing new Internet-based technologies.

Online services for mobile devices are designed using the end-to-end principle.

Data and services transactions are taken place between the service providers and

end-users. It is coupled with Internet-based technologies through certain standards,

such as Extensible Markup language (XML) [11], JaveScript Object Notation

(JSON) [12], and Simple Object Access Protocol (SOAP) [13] etc. In today‟s

market, the majority of “smart devices” are mostly based on Apple iOS [14] and

Google Android mobile Operating Systems (OS) [15]. They provide powerful

solutions for mobile platform to satisfy the growing requirements of mobile users.

Developers can freely design innovative mobile applications for mobile users.

Apple announced that there are over 75 billion downloads from the AppStore by

13
 Chapter 2. Identity, Privacy, and Security in Mobile Cloud Computing

June 2014 [16], and over 1.3 million apps available for iOS devices by September

2014 [17]. Meanwhile, Google Play reached 25 billion downloads with over

675,000 apps by 26th September 2012 [18].Google did not release further data on

the number of app downloads in the Play market. However, the Play market beats

AppStore with over 1 million apps in 2013 [19]. The AppBrain provides real-time

number of available apps in Google Play market [20], and there are 1432174

applications available in the Google Play market as of 17th December 2014. These

mobile apps give great convenience to people‟s life. Users are increasing the

demand of any sophisticated applications to do more work on their mobile devices.

The latest advances in mobile communication networks and the increasing

penetration of smartphones are transforming the mobile Internet and are

empowering end users with rich mobile experience [4]. However, the limited

onboard computing, energy and storage capabilities of mobile devices are

hampering the ability of smart devices to satisfy such demands. For instance, it is

impossible for a hotel reservation app to run at the smartphones as stand-alone

software. Because a user has to download a huge size database which stores all

related information, and the cost of operations such as search, update and reserve

from such apps are heavy. In order to address these limitations, Cloud computing is

identified as a possible solution [21].

Cloud computing has been widely recognized as the next generation‟s

computing infrastructure. It offers some advantages by allowing users to use

infrastructure, platforms, and software provided by Cloud servers at low cost [22,

23]. It represents a new paradigm shift in Internet-based service that delivers highly

scalable distributed computing platforms in which computational resources are

14
 Chapter 2. Identity, Privacy, and Security in Mobile Cloud Computing

offered “as a service”. In addition, Cloud computing enables users to elastically

utilize resources in on-demand fashion. As a result, mobile applications can be

rapidly provisioned and released with minimal management efforts or service

provider‟s interactions. With the explosion of mobile applications and the support

of Cloud computing for a variety of services for mobile users, mobile Cloud

computing is introduced as an integration of Cloud computing into the mobile

environment. Mobile Cloud computing brings new types of services and facilities

for mobile users to take full advantage of Cloud computing.

In Cloud environment, applications are delivered as services over the Internet.

The hardware and systems software in the Cloud data centres provide those services.

When a Cloud is made available in a pay-as-you-go manner to the public, it is

called a public Cloud; the service being sold is utility computing [24]. Currently

available public Clouds in the market include Amazon Elastic Compute Cloud

(EC2), Google AppEngine, and Microsoft Azure. The term of private Cloud refers

to internal data centres of a business entity or other organization that are not made

available to the public. Community Cloud is the Cloud infrastructure which is

shared by several organizations and supports a specific community that has shared

concerns. Hybrid Cloud is a composition of two or more Clouds (Private Cloud,

Public Cloud, or Community Cloud) that remain unique entities but are bound

together by standardized or proprietary technology that enables data and

application portability.

15
 Chapter 2. Identity, Privacy, and Security in Mobile Cloud Computing

The National Institute of Standards and Technology (NIST) defines fou4 types

of Cloud deployment models and three service models of Cloud computing [25]:

Cloud deployment models:

 Private Cloud: In this model, the Cloud infrastructure is provisioned

for exclusive use by a single organization comprising multiple

consumers. It may be owned, managed, and operated by the

organization.

 Community Cloud: The Cloud infrastructure is provisioned for

exclusive use by a specific community of consumers from

organizations that have shared concerns. It may be owned by one or

more organizations in the community.

 Public Cloud: The Cloud infrastructure is provisioned for open use

by general public.

 Hybrid Cloud: In a hybrid Cloud, the Cloud infrastructure is a

composition of two or more distinct Cloud infrastructures (private,

public, community) that remain unique entities, but are bound

together by standardized or proprietary technology that enables data

and application portability.

Service models of Cloud computing

 Cloud software as a Service (SaaS). The capability provided to the

consumer to run their application running on the service provider‟s

16
 Chapter 2. Identity, Privacy, and Security in Mobile Cloud Computing

Cloud infrastructure. The consumers are not required to manage or

control the Cloud infrastructure such as network, storage, or servers.

Users can access the services from various client devices through a

thin client interface such as a Web browser (web-based email). The

applications such as Gmail and Facebook are the popular mobile

Cloud computing SaaS applications.

 Cloud Platform as a Service(PaaS). Cloud consumers are able to

develop and deploy their own applications onto the Cloud

infrastructure using programming languages and tools supported by

the Cloud provider. Consumers do not required to manage the

underlying Cloud infrastructure but have the control over the

deployed applications and configurations of the application hosting

environment. The Amazon AWS is one of the emerging PaaS

product which could be used for mobile Cloud computing.

 Cloud Infrastructure as a Service (IaaS): The Cloud consumer has

the ability of processing, storage, networks, and other fundamental

computing resources where the consumer can deploy and run their

software including operating systems and applications. The

consumer does not manage the Cloud infrastructure but has control

over operating systems, storage, deployed applications and possibly

limited control of select networking components.

17
 Chapter 2. Identity, Privacy, and Security in Mobile Cloud Computing

Mobile Cloud computing aims at using Cloud computing techniques for

storage and processing of data on mobile devices to reduce their limitations. It is the

combination of mobile Internet and Cloud computing [26]. An illustrative example

of mobile Cloud computing is how a smartphone can best utilize the Cloud

resources to reduce its energy consumption. A computing task can be either

executed on the mobile devices or outsourced to the Cloud. Where to compute the

task is dependent on the overhead trade-offs between computation and

communication while considering the requirements of applications‟ Quality of

Service (QoS) and users‟ Quality of Experience (QoE) [27], where the QoS refers

to the overall performance seen by a customer [28] and the QoE refers to the

difference between expectations and perceptions for a service from a customer

[29].

Zhang and Yan proposed a QoS framework to manage QoS assurance in

mobile Cloud computing [30]. They used specific factors for mobile Cloud

services such as signal strength, mobile operating systems, hardware etc based on

the core factors such as reliability, assurance, tangibles, empathy, and

responsiveness. Researchers in [31] proposed a solution which is called

evidence-based method to measure QoE. They use six factors to provide the

evidences, which are review design, quality of review, consistency across services,

directness to services and business users of services, numbers of reviews, and

other related modifying factors.

18
 Chapter 2. Identity, Privacy, and Security in Mobile Cloud Computing

2.1.2 Advantages of Mobile Cloud Computing

Mobile Cloud computing improves the performance of many applications on

mobile devices. By using mobile Cloud computing, the following limitations of

mobile devices can be mitigated:

(1) Data storage capacity and processing power: Even though the hardware

specification of mobile devices is much improved, storage and

processing power are still the major constraints of the mobile devices.

For instance, the Instagram [32] service enables mobile users to upload

photos to Cloud immediately after capturing the images. Users can

browse all the photos from any supported devices. With mobile Cloud

computing, users can save energy and storage space of their mobile

devices. Such services are today successfully implemented by Flickr [33]

and Facebook [34] too.

(2) Battery life: Battery life is one of the most important concerns for

smartphones. With the modern hardware, mobile devices are equipped

with much more powerful processors, and more memory space to

process complex computations. However, these new hardware result in

an increase of energy cost and may not be feasible for all mobile devices.

Hence offloading complex computations to the Cloud can reduce the

operational cost of the mobile transactions. In addition it can also help to

conserve the battery life-time.

(3) Reliability: Cloud can provide more reliable and effective environment

to store data and hosting applications compared to the mobile devices.

19
 Chapter 2. Identity, Privacy, and Security in Mobile Cloud Computing

Factors such as the mobile operating system, battery, hardware etc. can

influence the reliability. Furthermore, the damage or theft of mobile

devices can cause unexpected loss if data are stored locally. By using

Cloud, such loss can be minimal since the data are stored and processed

on the Cloud.

(4) Flexibility and Scalability: Cloud resources can be consumed in an

on-demand, fine-grained, and self-service manner. It is more flexible for

service providers and mobile users to run applications without

reservation of resources. The deployment of applications can be scaled

to satisfy the changes for demands of users, and service providers can

easily update their services without considering resource usages.

(5) Ease of integration: Mobile users can consume services from different

service providers since it is easy to integrate multiple services in the

Cloud. An example of multiple services integration is that of

Instagram users who can share their photos on Facebook or Flickr if

they authorize Instagram with access to their personal data and page.

Due to these advantages, the mobile Cloud computing is largely adopted in

Mobile Commerce, Mobile Health, Mobile Education and Mobile Gaming

applications. Such mobile applications are developed in a global mobile market.

Gartner‟s report says that by 2016, nearly 40% of mobile application development

project will leverage Cloud mobile back-end services [35]. However, there are

several drawbacks in mobile Cloud computing that needs immediate attention

20
 Chapter 2. Identity, Privacy, and Security in Mobile Cloud Computing

before it can take off in commercial arena. The drawbacks associated with mobile

Cloud computing are discussed in the next section.

2.1.3 Issues of Mobile Cloud Computing

With the advantages discussed above, different types of mobile Cloud computing

services greatly increase convenience, however, new issues are introduced which

may affect the QoS and QoE of mobile Cloud computing.

(1) Network bandwidth and latency: Bandwidth and communication

latency are big issues in mobile Cloud computing since the radio

resource for wireless networks is much scarce as compared with

traditional wired networks. Bandwidth for 3G cellular systems may

be limited by cell tower bandwidth in some areas with low power

signal receptions which in turn leads to lower bandwidth and higher

latency [26]. The incoming 4G wireless network can address this

issue. Wi-Fi is also considered as a solution, but the performance of

Wi-Fi depends on the number of mobile users. If there are a large

number of users, the bandwidth will also be decreased.

(2) Network availability and intermittency: Constant and speedy

Internet connection must be ensured in mobile Cloud Computing.

Mobile users may not be able to connect to the Cloud to obtain

service due to traffic congestions, network failures and loss of signal.

In [36], authors proposed a Wi-Fi based multi-hop networking

system to address this problem. A list of neighbour nodes is selected

by a user, and if there is a loss of connection, the nodes with highest

21
 Chapter 2. Identity, Privacy, and Security in Mobile Cloud Computing

weight value will continue to receive content. The new Hyper Text

Markup Language (HTML) 5 comes with a function of data caching

through mobile devices. It is also possible for mobile Cloud

application to address this issue.

(3) Resource poverty of mobile devices: The limited onboard computing

power and storage spaces were the major issues of mobile devices.

With the development of hardware, such problems can be resolved

so that the mobile devices are able to display as much information as

required to obtain maximum convenience.

(4) Security concerns: The advancement in technology has also brought

new security challenges. Everyone wants to protect their personal

data online. Protecting user privacy and data secrecy from an

adversary is a key to establish trust in mobile Cloud environments.

The following two aspects of security related issues in mobile Cloud

environment should be considered:

a. Security for mobile user: Mobile users utilize Clouds for

computing and storage resources. Client applications may be

required to be installed on the mobile devices. The security

of the running environment (i.e. operating system) of client

applications should be guaranteed and protected from threats,

such as malware, viruses, and Trojans. Meanwhile, the

design of client application should authenticate the user and

protect it from being used by unauthorized users. On the

22
 Chapter 2. Identity, Privacy, and Security in Mobile Cloud Computing

other hand, mobile Cloud applications will request user‟s

real-time context information, such as location-based

services. Mobile users face privacy issues as service

providers may use the important personal information for

marketing purposes without prior permission.

b. Security at Cloud: Mobile users benefit from storing data

and hosting applications on the Cloud. Data integrity and

confidentiality on the Cloud is one of the major concerns. It

is important that no unauthorized users can reveal the data.

Secondly, it is necessary to have access control over all of

the user data on the Cloud. Thus, a user can decide which

data requester can obtain access and know the purpose of the

data usage. Thirdly, different service providers may holds

partial identity information of a user. Avoiding collusion

attack in mobile Cloud environment is also an important

issue.

The above issues are main concerns of the mobile Cloud environments. The

main focus of this thesis is on the data security and privacy issues in the mobile

Cloud environment and the protection of users‟ sensitive personal information such

as identities and identity related attributes during electronic transactions. In the

following sections, a case study for the investigation of current relationships within

identity, privacy, and security technologies in presented in the context of mobile

23
 Chapter 2. Identity, Privacy, and Security in Mobile Cloud Computing

Cloud Computing in Section 2.2. The current security primitives to protect the user

identity and privacy in Section 2.3.

2.2 Case Study

Recall the example used in Chapter 1. If an Instagram user wants to share his photos

with his friends on Facebook. He needs to authorize Instagram using his personal

Facebook account details. Figure 2.1 shows the default authorization settings of

Instagram at a Facebook account. Instagram will ask for the permission to contents

on Facebook page and access to data at any time. As a result, the user‟s identity

information and attributes recorded at Facebook are revealed to the Instagram

service. This is a typical identity information authorization scenario in a mobile

Cloud environment.

A mobile Cloud user registers at service provider with a unique identity. For

example, a user can use his email address as a unique identity to register at

Facebook. The email identity may consist of a set of attributes such as surname,

date of birth and place of birth to describe the user. These attributes can be either

static or real-time. Attributes such as username, home address, date of birth, and

device IMEI number can be considered as static attributes which will not be

changed frequently. The identity in mobile Cloud environment normally includes

real-time contextual attributes such as users‟ location records, time, and nearby

devices. These contextual attributes can describe privacy information of a user. It is

a key challenge to protect privacy while securing users‟ digital identities in the

mobile Cloud environments [37].

24
 Chapter 2. Identity, Privacy, and Security in Mobile Cloud Computing

It raises an identity security issue that if two such service providers collude

together, then they can reveal all the user‟s attributes. Existing identity

management systems, such as OpenID [38], Higgins [39] etc. are designed for

traditional PC-based Cloud environment, and do not work with all their features in

mobile Cloud environments. Such identity management systems will be discussed

in detail in Chapter 3.

Figure 2.1Authorization Information of Instagram at a Facebook Account.

In today‟s mobile Cloud environment, mobile apps deliver convenient services

to the user with good experience. This trend increasingly affects the user privacy.

Let us take the location-based service as an example. Location-based applications

are typical mobile Cloud applications. A user wants to book a table at a restaurant

with best reviews around his current location. He uses a mobile app which has a

large number of restaurants information stored in the Cloud database, such as

Tripadvisor [40]. User‟s mobile device captures the current location and sends it to

25
 Chapter 2. Identity, Privacy, and Security in Mobile Cloud Computing

the Cloud service provider hosting the mobile app. The app analyses the location

and returns the requested results. In order to complete the booking, the user needs to

give more details about him to the app. During this process, user does not know

whether such personal details are collected by the app or some other third party

entities. Furthermore, with a long time usage, the app may have a huge number of

data based on the user‟s behaviours. The user will not know how his personal

information will be used and even does not know his data is collected by the mobile

Cloud service provider.

Most of the service providers claim that they will only collect anonymous data.

However, when the service provider has a large number of users, it may sell the

users‟ sensitive data to a third party who can mine the data to build business models

for profits [41-43]. Mobile apps can also be used to inject the malwares to users‟

phones. Some malwares can even monitor user‟s daily activities [44]. Researchers

have investigated this issue and proposed solutions in order to prevent collusion

attacks. Bugiel et al. presented a practical security framework for Android platform

which addresses the problem of collusion attacks [45]. Their framework can

monitor application communication channels in Android's middleware. A

system-centric security policy-based mechanism is deployed to enforce access

control in Android platform. They also improved the security framework with a

system kernel module [46], which enables the framework to provide security

protection at system level.

The problem of identity security and privacy is to find a balance between user

convenience and computer security. The best way is not to give any information in

26
 Chapter 2. Identity, Privacy, and Security in Mobile Cloud Computing

order to protect privacy. However, in many cases users are often asked to give some

personal information during system authentication process to gain access to data

and services. Researchers have also put efforts to address issues raised above.

Various access control techniques such as privacy-preserving languages, access

control models, and attribute-based encryption are deployed to protect users' data.

Chapter 3 and Chapter 4 details those technologies and gives a literature review

for related work.

2.2.1 Existing Data Protection Laws

Privacy protection must control what information about an individual is disclosed

over the Internet. It includes two aspects: 1) the users should have the control of

giving data access consents about their personal information to any authorized

entities (i.e. it determines who can access the data) and 2) the control of how the

data can be used (i.e. it determines the purpose of data usage).

The Organisation for Economic Co-operation and Development (OECD) also

published guidelines on the protection of privacy and trans border flows of personal

data [47]. The guidelines have been used to derive privacy laws governing the use

of information systems in a number of countries. It introduces eight basic principles

for data privacy protection which are listed below.

 Collection Limitation Principle: The collection of personal data and

any such data should have limits. And the collection means should be

lawful and fair, and, where appropriate, with the knowledge or consent

of the data subject.

27
 Chapter 2. Identity, Privacy, and Security in Mobile Cloud Computing

 Data Quality Principle: Personal data should be relevant to the purpose

of use, and the extent for those purposes should be accurate, complete

and kept up-to-date.

 Purpose Specification Principle: The purpose for personal data

collection should be specified earlier than the time of data collection and

the subsequent use limited to the fulfilment of those purpose or such

others as are not incompatible with those purpose and as are specified on

each occasion of change of purpose.

 Use Limitation Principle: Personal data should not be disclosed, made

available or otherwise used for the purpose other than those specified at

the collection time, except with the consent of the data subject or by the

authority of law.

 Security Safeguards Principle: Reasonable security safeguards should

be deployed to protect personal data against such risks as loss or

unauthorised access, destruction, use, modification or disclosure of data.

 Openness Principle: There should be a general policy of openness about

developments, practices and policies with respect to personal data.

Means should be readily available for establishing the existence and

nature of personal data, and the main purpose of their user, as well as the

identity and usual residence of the data controller.

 Individual Participation Principle: An individual should have the right:

28
 Chapter 2. Identity, Privacy, and Security in Mobile Cloud Computing

a) To obtain from a data controller, or otherwise, confirmation

of whether or not the data controller has data relating to him;

b) To have communicated to him, data relating to him

I. within a reasonable time;

II. at a charge, if any, that is not excessive;

III. in a reasonable manner;

IV. in a form that is readily intelligible to them;

c) To be given reasons if a request made under subparagraph

(a) and (b) is denied, and to be able to challenge such denial;

d) To challenge data relating to him and, if the challenge is

successful to have the data erased, rectified, completed or

amended.

 Accountability Principle: A data controller should be accountable

for complying with measures which give effect to the principles

stated above.

The USA government published the Health Insurance Portability and

Accountability Act (HIPAA) in 1996 [48]. The act addresses the security and

privacy issues of health data. The European Union (EU) published the Data

Protection Directive which regulates the processing of personal data within the EU.

It defines that personal data can only be gathered legally under strict conditions for

a legitimate purpose. Organizations who collect and manage user‟s personal

information must protect it from misuse and must respect certain rights of the data

29
 Chapter 2. Identity, Privacy, and Security in Mobile Cloud Computing

owners which are given in the EU law [49, 50]. In 2013, the European Commission

proposed a major reform of the EU legal framework on the protection of personal

data which is aimed at strengthening individual rights and tackling the challenges of

globalisation and new technologies.

2.3 Security Concepts, Technologies and Mechanisms

This section examines the current technology building blocks and standards for

developing security frameworks in identity management. In mobile Cloud

environment, communications and data exchanges of mobile devices are taking

place via wireless medium. Similar to Ethernet, three key aspects of information

security referred to as CIA triad: Confidentiality, Integrity and Availability [51]

must be followed in wireless communication. Additional aspects to security such as

Non-repudiation, Accountability are also referred to traditional CIA triad [52].

The key aspects are described below.

Confidentiality refers to the data concealment, preventing the disclosure of

information to unauthorized entities.

Integrity protects information resources form modification and deletion by

unauthorized parties. It ensures the accuracy and consistency of information

resources over the life-cycle.

Availability serves the purpose that the information resources must be

accessible when it is required. It ensures that the system work promptly and the

services are not denied to authorized users.

30
 Chapter 2. Identity, Privacy, and Security in Mobile Cloud Computing

Non-repudiation implies the occurrence of an event or an action and the

participant. A party cannot deny receiving a request and the other party cannot deny

that it is the sender of the request.

Accountability implies the actions or decisions which are made by an entity

are accountable.

The concepts listed above are necessary requirements for any security

frameworks. These concepts help to protect the information systems from attacks.

Security mechanisms are normally designed based on those concepts.

2.3.1 Security Technologies and Mechanisms

Information systems implement security mechanisms to detect, prevent or recover

from a malicious attack. It consists of series of security services such as

authentication, authorization and access control etc. Various security techniques are

used for these services such as cryptography, digital signatures [53], Transport

Layer Security(TLS) [54] protocol, and Security Assertion Markup Language

(SAML) [13].

Cryptography is the mathematical approach to secure communication in the

presence of adversaries. It will satisfy confidentiality of messages. The information

should not be revealed to any other parties except the intended recipient. Prior to the

modern age, cryptography is synonymous with encryption. The encryption is one of

the important areas in cryptographic technologies. The message which contains the

fully readable information is called plaintext. The encryption process is used for

31
 Chapter 2. Identity, Privacy, and Security in Mobile Cloud Computing

scrambling or disguising the plaintext by applying cryptographic algorithms. A

ciphertext is generated based on the plaintext after encryption process. The

decryption process converts the ciphertext to plaintext. It is the reverse of the

encryption process. A key is required for both encryption and decryption to perform

cryptographic algorithms. It is a secret value which is sensitive and should not be

revealed to any other third party. The key normally is a special numerical value

which is long and large enough to prevent from attacks.

The current cryptography technologies can be classified by two main

categories: the secret key technology and the public key technology. In the secret

key cryptography, the sender and receiver use the same key for encryption and

decryption. Therefore, the key is shared as a secret by the two ends but is blind to

the outer world. Different from the secret key cryptography, the public key

cryptography has a pair of keys, a private key and a public key. The public key is

open to any one while the private key is only known to the key holder. Either key of

the key pair can be applied for encryption, but only the other key of the same key

pair can do the decryption. Based on these two main technologies, several known

cryptographic mechanisms are described in the following.

2.3.1.1 Symmetric Cryptography

Symmetric cryptography is the oldest and best-known cryptographic technique. It

also refers to the secret key cryptography. The secret key can be a number, a word

or a string of random letters. The information system takes as input a secret key to

the encryption algorithm with the plaintext to generate the ciphertext. Figure 2.2

32
 Chapter 2. Identity, Privacy, and Security in Mobile Cloud Computing

depicts the work flow of the symmetric cryptographic mechanism. There are two

commonly used types of symmetric encryption algorithms that are listed below:

Block ciphers takes as input a block of data and encrypt them as a single unit

with a key. Blocks of 64 or 128 bits have been widely used. Such algorithms are

implemented in a variety of ways. For example, the Advanced Encryption Standard

(AES) [55] and the triple-DES (Data Encryption Standard) [56].

Stream ciphers encrypt the digits (typically bytes) of a message one at a time.

The stream cipher algorithm proposed in [57] is a widely used model.

Figure 2.2Symmetric En/Decryption Process

2.3.1.2 Asymmetric Cryptography

In symmetric cryptography, securely exchanging the secret keys over the Internet

or public domains is a critical issue. Users have to prevent from sending them to

wrong recipients. Asymmetric cryptography is a solution. It is also known to as

33
 Chapter 2. Identity, Privacy, and Security in Mobile Cloud Computing

public key cryptography. Instead of using pre-shared secret keys, every user must

have a key pair. The key pair consists of two keys: a private key and a public key.

Both the private and public keys are generated or acquired by owner of the key

pair. The public key is publicly available and can be used for encryption,

verification of a digital signature (digital signature will be discussed in the next sub

section). The private key is secured at the owner and is kept as a secret. In a public

key scenario, a sender encrypts a message with the recipient‟s public key. The

receiver who holds the corresponding private key can decrypt the message. Figure

2.3 describes work flow of the asymmetric cryptographic mechanism.

There are several existing asymmetric encryption algorithms, the EI Gamal [58]

and Rivest-Shamir-Adleman (RSA) [59] are the most widely used public key

schemes.

Figure 2.3Asymmetric En/Decryption process

34
 Chapter 2. Identity, Privacy, and Security in Mobile Cloud Computing

2.3.1.3 Digital Signature

Digital signature supports the non-repudiation aspect of the information exchange.

In digital signature scheme, a message is signed by the sender using the private key.

The recipient can verify the signature using the sender‟s public key. The digital

signature scheme gives an evidence for a recipient to believe that the message was

sent by a known sender. The sender cannot deny the fact of having sent a message.

It is used to preserve non-repudiation and authentication. The Digital Signature

Algorithms [60] (DSA) and RSA [53]are commonly used.

2.3.1.4 Hash Function

Hash function is used to provide integrity of messages. A hash function is a public

function which maps any length of data input into a fixed length output. Given a

message M of arbitrary length, a hash function H will produce a fixed-sized hash

value h in a form of h = H(M). h is a hash value of message digest. The properties

of a hash function [61] are listed as follows:

One-way property: H(x) is easy to compute h with any given input x, however

it is computationally infeasible to compute such that H(x) = h.

Weak collision resistance: For any given x, it is computationally infeasible to

find y ≠ x with H(y) = H(x).

Strong collision resistance: It is hard to find any pair (x, y) to compute such that

H(x) = H(y).

By applying a hash function, the secret parameters such as password, secret

keys can be verified based on the hash values of them. It greatly helps for data

35
 Chapter 2. Identity, Privacy, and Security in Mobile Cloud Computing

validation with data privacy. Secure Hash Algorithm (SHA) is a hash algorithm and

it was developed in 1993. The cryptography hash algorithm SHA-1 was published

in 1995 with a maximum input length 264 bits and outputs a 160 bit message

digest[62]. However, a successful attack on SHA-1 was reported in 2005 [63] and

in order to avoid these problems, more advanced algorithms such as SHA-2 and

SHA-3 in SHA family are used in modern day secure communication.

2.3.2 Protocols and Standards

2.3.2.1 Transport Layer Security (TLS)

The Transport Layer Security and its predecessor, Secure Sockets Layer (SSL), are

both cryptographic protocols which provide a channel communication security

over the Internet [64]. It secures the communication channels above the IP layer and

below the application layer. In a TLS/SSL scenario, asymmetric cryptography is

used for authentication of key exchange. Symmetric cryptography is implemented

for data exchange.

SSL was first developed by Netscape [65], it enables a client-server application

(normally is a Web browser and a Web server) to exchange data over public

domains. Data confidentiality and integrity are ensured. The latest version of TLS is

published in 2008 [54].

2.3.2.2 eXtensibleMarkup Language (XML):

eXtensibleMarkup Language (XML) is a self-descriptive, platform-independent

markup language. It is free, open standard and resembles a general-purpose version

of HTML. XML is designed to represent, transport and/or store a range of

36
 Chapter 2. Identity, Privacy, and Security in Mobile Cloud Computing

structured data. It enables users to define their own tags in XML document and is

widely used for business data share and exchanging in different domains. XML

parser is a software program that checks the syntax of the XML document and

makes the XML document‟s data readable to applications. XML schema is a

method of describing the semantic rules for a XML document to be validated. A

XML parser firstly checks whether the document is well formed, the term of

well-formed means satisfy all XML syntax rules. Then the parser will check

whether the document is Valid, which means satisfied semantic rules that are in

XML schemas. W3C published the XML specification in 2008 [48].

2.3.2.3 Simple Object Access Protocol (SOAP)

SOAP is a standard lightweight transport protocol in Web services used for

exchanging structured information in a decentralized, distributed environments

over HTTP protocol. SOAP specification defines, using XML technologies, an

extensible messaging framework containing a message construct that can be

exchanged over a variety of underlying protocols [13]. In SOAP message systems,

different applications can interoperate by negotiating both parties into a common

data transfer protocol. Typically, a simple SOAP message is enclosed inside an

SOAP envelope. The envelope contains two main regions such as the SOAP

Header and SOAP Body. The SOAP Header contains information about the

SOAP message and the SOAP Body contains the message payload.

37
 Chapter 2. Identity, Privacy, and Security in Mobile Cloud Computing

2.3.2.4 Security Assertion Markup Language (SAML)

SAML is an XML-based framework that allows the exchange of authentication and

authorization information between two Web service domains. There are four key

elements covered by the SAML specification.

SAML Assertions: a pack of XML-based data includes one or more statements

made by a SAML authority.

SAML Protocols: Defines data structures of request and response messages on

how to obtain assertions.

SAML Bindings: Map SAML protocol messages onto standard communication

protocols, e.g. HTTP.

SAML Profiles: Describe particular combination of SAML assertions,

protocols and binding for a particular user case.

SAML 1.0 was first adopted as a Web standard by OASIS in 2002 [66]. After

several revisions, in 2005, SAML 2.0 was announced [67]. It has been widely

adopted by government, higher education and commercial enterprises.

SAML protocol messages are transported within the SOAP body and SAML

response has SAML status and one or more assertions. The assertion consists of one

or more statements which contains a single authentication statement. SAML is

widely used in federated environment because it enables the functionality that users

can authenticate themselves in one Web service domain without re-authentication

in another domain.

38
2.4 Summary

The mobile Cloud technology is a combination of Cloud computing and mobile

Internet technologies. In general, when a user interacts with a service provider,

some static user attributes may be gathered by the service provider. However, in

mobile Cloud environment, not only static attributes but also dynamic user

attributes such as location records and user behavior are collected by the service

providers. Hence, this new notion adds various new threats to the user privacy.

Identity management systems manage users‟ identities; it is also necessary to

modify the security concepts to cope with the emerging mobile application trend. In

particular, the current security concepts used in identity management is powerful

enough for securing data and service transactions. Even though, smart mobile

devices can use these security features, these is a necessary to develop new

techniques to protect the privacy of real-time dynamic attributes in mobile Cloud

environments.

Due to the nature of mobile Cloud environments, service providers are now

collecting much more sensitive information about the user with or without their

consents. Users should have full control over their own data. Security technologies

such as access control, and cryptography should enable the users to disclose

personal data based on their preferences. The existing technologies are able to

secure data transaction and confidentiality, but they cannot cover such new privacy

39
 Chapter 2. Identity, Privacy, and Security in Mobile Cloud Computing

issues in today‟s mobile Cloud environment. Besides security, privacy protection

has become a critical issue in modern day mobile applications.

40
3 Identity Management Systems

As the number of mobile Cloud services requiring user authentication continues to

grow, so does the number of digital identities. To help protect digital identities, a

range of identity management systems have been proposed to address security

issues of identity protection and identity management. These proposed systems are

designed to simplify the management of a large number of identities, meanwhile,

protect identities against cyber-attacks. This chapter briefly surveys various

identity management systems in the literature.

This chapter is organised as follows. The concept of Single-Sign-On (SSO)

and federation which are implemented today by most of the identity management

systems is introduced in Section 3.1. In Sections 3.2-3.6 describe .NET Passport,

the Liberty Alliance, OpenID, Higgins, and OAuth respectively. A discussion of

the above identity management systems and literature review is given in Section

3.7. Finally, the Chapter is concluded in Section 3.8.

41
 Chapter 3. Identity Management Systems

3.1 Single-Sign-On(SSO) and Federation

Different online services may be hosted by different service providers. Users have

to do the authentication to each service provider separately in order to access the

different the Cloud services. SSO is a technology that lets users authenticate to a

single authentication authority once so that they can access all authentication

protected resources and services without re-authentication within the federated

environment. Therefore, one login action enables user access to all the permitted

resources in a complex system without entering multiple usernames and passwords.

SSO provides convenience to both the user and service provide. Users do not have

to memorize many usernames and passwords and at the same time, it also reduces

the number of authentication requests to the service providers.

Federation, or Federation of identity, describes the technologies, standards

which serve to enable identities to obtain access across autonomous security

domains. A user of one domain can securely access data or services seamlessly

from networks of other domains. SSO is a subset of the federated identity

management framework. The user authenticates to the authentication server by

submitting credentials such as username, X.509 certificate and identity related

attributes etc. The credential includes the user identity and security information.

The user identity is known as a federated identity. It is used in order to permit the

user to gain access across different domains which are covered by the same

authentication server using the same identity. The authentication server is named as

42
 Chapter 3. Identity Management Systems

an identity provider. It takes charge of the authentication and authorization and

maps the federated identity to other user identity credentials.

SSO not only provides convenience to traditional PC users in the past, but also

can continue to play an important role in today‟s mobile environment. The

simplification of identity management delivers excellent user experience in mobile

environment. Identity management systems satisfy requirements for security and

user experience in different mobile environment. There are several identity

management systems in today's market. In the following sections, these identity

management systems are briefly described in detail.

3.2 Microsoft .NET Passport:

In 1999 .NET Passport was introduced by Microsoft [68]. In a .NET Passport

system, the users can access multiple websites using a single set of credentials.

Microsoft presents .NET Passport system as their web-based SSO solution. A

Passport user gives permission to the Passport server to the Passport account

credentials. The credentials consist of username, password, email address, first

name, last name, date of birth, etc. The sites deploying Passport services are called

participating sites. Microsoft provides Passport services as a centralized

authentication server.

.NET Passport frees the user from multiple registrations and memorizing

passwords. The following figure explains how the Passport system works.

43
 Chapter 3. Identity Management Systems

Figure 3.1.NET Passport Authentication Process

(1) User requests resources from a Passport participating site via the

browser

(2) The participating site generates a Passport object and forward to the

browser to check whether there are any valid passport cookies

(3) If there are any existing cookies, the page loads with a sign-out link and

consider the user as an authenticated user. Otherwise, the page loads

with a sign-in link.

(4) The user clicks the sign-in link to start the authentication process. The

page then redirects to the .NET Passport sign-in page. The user‟s

browser sends the Passport authentication request to the .NET passport

server. The request contains the Passport Manager Object, the

44
 Chapter 3. Identity Management Systems

participating site ID, the return URL to the participating site after

authentication

(5) .NET Passport Server checks the validity of the participating site ID

and returns the URL and the Passport sign-in page

(6) If the user is successfully authenticated by the .NET Passport server.

The server generates encrypted cookies based on user‟s Passport User

ID and the Passport profile and then forwards these data to the user‟s

browser with the redirection URL. Cookies are encrypted by the

participating site‟s public key. The communication is secured by

TLS/SSL channel encryption

(7) The browser redirects the page to the participating site with the

encrypted cookies

(8) The Passport Manager authenticate the cookies and grant access to the

user for accessing the site

.NET Passport is a SSO mechanism for systems that are connected over the

Internet. User authentication is implemented via web forms. Although today‟s

mobile browsers, such as Safari and Chrome are designed with many important

functionality to execute operations as a traditional PC browser, the .NET Passport is

not a suitable for mobile Cloud environment because of its browser-based feature

and lack of participation from users.

In 2008, Microsoft announced that .NET Passport becomes an identity

provider of the OpenID framework [69]. The users of .NET Passport system are

allowed to access any web sites that are covered by the OpenID authentication.

45
 Chapter 3. Identity Management Systems

After this planned implementation in 2009, there is no further update for .NET

Passport [70].

3.3 The Liberty Alliance(Kantara)

The Liberty Alliance is a large consortium which is established in 2001 by

approximately 30 organizations. Now it has over 150 diverse member companies

and organizations representing leaders in IT, government, and finance from across

globe to build open, standards based specifications for federated identity

management. The vision of the Liberty alliance is to establish privacy and security

best practices as well as implementation guidelines. It also collaborates with other

standards bodies with an eye toward adopting or extending other specifications [71].

From the mid of 2009, the work of the Liberty Alliance project has been to

contribute to the Kantara Initiative [72]. Figure 3.2 depict the abstract view of the

Liberty architecture.

Figure 3.2 The Abstract View of Liberty Architecture

46
 Chapter 3. Identity Management Systems

The Liberty Alliance project consists of a number of modules. Each module

aims to solve different problems in identity management. The following section

will discusses them in detail.

Liberty Identity Federation Framework (ID-FF): The Id-FF enables identity

federation and management. A Liberty Federation allows users to authenticate and

sign-on to a domain once from any device and then access to the partner Web sites

without re-authenticate. It is independent of platform, network devices, computing

and mobile devices [73]. It provides approaches for implementing federation and

SSO.

Liberty Identity Web Services Framework (ID-WSF)[74]: The ID-WSF is an

open framework for deploying and managing a variety of identity-based Web

services. It contains a set of specifications for creating, using and updating various

aspects of identities. The ID-WSF also presents security functions for privacy

protection.

Liberty Identity Services Interfaces Specification (ID-SIS): The ID-SIS is an

open framework which contains a collection of specifications. These specifications

are designed for the interoperability of different services [75].

Liberty Alliance is based on the notation of „trust circles‟ which are formed

by a trusted authentication server and set of service providers. The trust

relationship of the authentication server and service providers are supported by

contractual agreements and it is outside the scope of Liberty Alliance

specifications. According to the specification, the user authenticates to the

authentication server and it expresses the authentication assertions to the relying

47
 Chapter 3. Identity Management Systems

service providers. The assertion consists of the name identifier of the user for

service providers to identify the user. The authentication server uses distinct

identifiers for each user with different service providers. Figure 3.3 describes the

Liberty notation of Trust.

Figure 3.3The Liberty Notation of Trust

The Liberty Alliance supports SSO and use SAML token for authentication.

Using SAML also enable it to bring attributes within the communication messages.

However, it does not support attribute exchange which is critical in mobile Cloud

environments.

3.4 OpenID

OpenID is an open source and decentralized identity management system that

supports Web SSO to multiple sites using a single digital identity [76]. The OpenID

framework was announced in 2005. An OpenID user can create an account with a

preferred OpenID identity provider and then use the account for signing into any

sites which accepts OpenID authentication. There are three main actors under the

48
 Chapter 3. Identity Management Systems

OpenID framework: the user, the Relying Party (RP), and the Identity Provider

(IdP). The RP is always known as the service provider from which the user requests

services.

In the OpenID framework, three parties are involved in the authentication

process. They are theOpenID provider (OP), which stands for the identity provider;

the Relying Party (RP) which is also called service provider and the user. Assuming

that the OP and the RP has an existing trust relationship. OP has a trusted list of RPs.

In the OpenID model, the users only need to have the identity and password details.

Figure 3.4 describes the workflow of OpenID framework.

Figure 3.4OpenID Authentication Process

49
 Chapter 3. Identity Management Systems

(1) User requests to login into the RP and submits the credential

(2) RP discovers the OP information from the user‟s credential and

redirects the OP to the user

(3) User enters the password into the field on the OP‟s page

(4) If the user‟s credential is verified, OP redirects RP‟s page to the user

OpenID users can choose a trustworthy OpenID server to register their OpenID

account. They are identified as a URL: http://yourname.openidserver.com. The RP

parses the URL and will get the OP‟s address “openidserver.com” and the user‟s

identity “yourname”. Then the RP can redirect the OP to the user for authentication.

The authentication process will make use of the identity “yourname” and the user

will inputs the password. If the authentication is successful, OP will redirect RP‟s

page to the user. In the whole process, user is not required to reveal the password to

the RP which protects the privacy [77]. However, limitations are discussed in [78]

that OP only uses one password to authenticate the user which is not secure enough.

OpenID is a symmetric cryptography based framework, which is simple to use but

more vulnerable than asymmetric encryption. Moreover, it is a web browser based

mechanism and it utilizes the cookies which is not suitable for the mobile Cloud

environment.

There are two main versions of OpenID specification: OpenID 1.1 [79] and

OpenID 2.0 [38]. The 2.0 version is compatible with the 1.1 version. OpenID has

been largely adopted with one billion accounts and nine million sites enabling

50
 Chapter 3. Identity Management Systems

OpenID authentication services [80]. IT enterprises like Google, Yahoo, and

Microsoft are acting as the OpenID Identity Providers.

3.5 Higgins

Higgins [39] launched in 2003, an open-source Internet identity framework

designed to integrate identity, profile, and social relationship information across

multiple sites, applications and devices. Higgins is InfoCard-based and involves an

active client.

Higgins provides a Personal Data Service (PDS) that let each user control how

their personal data is shared with friends and organizations they trust. A PDS is a

Cloud-based service that works on behalf of a user, the individual. It gives the user

a central point of control for his personal information. The user‟s interests, contact

information, address, profiles, affiliations, friends etc. can be considered as user

personal data. A PDS is a place where a user establishes bi-directional data flows

between external businesses and his PDS, or between a trust recipient‟s PDS and

his PDS. Figure 3.5 depicts the architecture of Higgins 2.0.

Higgins uses the concept of cards to store information between the user and the

other entity. The entity can be either a service provider or a person. In Figure 3.5,

the cards shown in green represent a relationship between the user and an external

site or business. The relationship includes a bi-directional data connection that

shares and synchronizes a set of attributes between the site and the user‟s PDS.

51
 Chapter 3. Identity Management Systems

Figure 3.5Architecture of Higgins

The profile shown in brown circle represents a relationship between the user

and another person. The relationship includes a bi-directional data connection that

shares and synchronizes a set of attributes representing a friend.

The blue box shown as AddOn App is a built-in app or extension app which has

access to the local storage and update, add value to what‟s available irrespective of

any particular relationship connections.

The local storage for data storage holds user‟s data using the vocabularies of

the Higgins‟ Persona Data Model 2.0. It exposes these data to the Portal or a

browser extension via an HTTP/Comet messaging interface.

Higgins provides online identity management services in a profile/card based

manner. A card is established for the user to control the data release to a preferred

recipient. However, there are some limitations of Higgins. Although the client

52
 Chapter 3. Identity Management Systems

apps are design as browser-based or a stand-alone client app, they cannot run

perfectly on mobile devices at the moment. This is a major drawback in mobile

Cloud environment.

3.6 OAuth

OAuth is an emerging, open, identity management standard. It describes a system

which is designed to enable an end user to grant an Internet application controlled

access to personal information (e.g. user attributes, photos, location records) stored

at a third party site, without divulging long-term credentials such as passwords.

Briefly, it allows a user, to grant access to your private resources on one site to

another site. While OpenID in Section 2.6.2 is about using a single identity to sign

into many sites, OAuth is about giving access to your stuff without sharing your

identity at all [81].

There are four entities involved in the OAuth protocol:

(1) Resource Owner: refers to an end user of a hosting site.

(2) Client: refers to the application (requesting site) to user‟s resources.

(3) Resource Server: refers to the hosting site of user‟s resources

(4) Authorization Server: refers to the server that issues an access token

to a client after successfully authenticating the user and obtaining its

authorization.

53
 Chapter 3. Identity Management Systems

OAuth 2.0 [82] was published in 2011 and many IT leading enterprises such as

Facebook, Google, LinkedIn are on the list of OAuth service providers. Recall the

Instagram scenario, the OAuth protocol enables an Instagram user authorizing on

Instagram access to personal data of his Facebook account. It is easy to use and

saves plenty of time for the user. An OAuth user can freely choose which service

provider can obtain the authorization to his personal content from the hosting site,

but also define which part of data and time period that the requesting service

provider can access. Your identity credentials such as password and username will

not be revealed to the requesting services party. Hence user‟s privacy is partially

protected.

The following Figure 3.6 depicts the work flow of the OAuth protocol.

Figure 3.6Workflow of OAuth Protocol

54
 Chapter 3. Identity Management Systems

Although, OAuth limits new sites from obtaining all of user data and selling it

against your will, it cannot prevent abuse of user‟s personal data. One of the largest

OAuth implementation is Facebook which shares user‟s data with business partners.

Those third-party applications in the App Store of Facebook have the rights to

request access to personal data such as a user‟s friendship details, email address,

education history, hometown etc. These settings make the user‟s privacy under

threat. If one or more service providers combine their data together, the user will be

face with the possibility of collusion attack.

3.7 Comparison and Literature Review

A generic comparison of the above identity management systems is given.

Table 3.1 gives the results of comparison, where √refers to supported and ×refers

to not supported.

Attribute SAML
Type SSO
Exchange Tokens

.NET
Redirect-based √ √ ×
Passport

The Liberty
Redirect-based √ × √
Alliance

OpenID Redirect-based √ √ ×

55
 Chapter 3. Identity Management Systems

Higgins Client-based √ √ √

OAuth Redirect-based × √ ×

Table 3.1 Comparison of the Different Identity Management Schemes

An identity management system should be able to protect private and sensitive

information related to users and process [83]. The private and sensitive data not

only contain identity information, but also contain privacy information. The

traditional identity management technologies provided by existing identity

management systems should be improved to support the emerging mobile Cloud

technology. Researchers have done a lot of work in order to improve privacy

protection of current identity management systems.

Sun and Yan proposed a generic and flexible solution based on Cloud computing

technology [84].They built a Cloud-based trusted identity management system. The

system carries out authentication and authorization among heterogeneous mobile

networks. They solved the WLAN/Cellular integration issues using Cloud-based

technologies. However, they did not consider the privacy protection. Authors [85]

presented a framework for mobile identity management. They used strong

authentication methods to provide solutions encompass user access, signing and

verification of users and transactions. However, strong authentication requires

users to do further actions based on the requests, and this framework did not

consider the user data protection.

56
 Chapter 3. Identity Management Systems

For a smartphone user, the GPS sensor can capture the real-time location data, the

camera can be used to recognize a person's face etc. Thus, lots of contextual data

can be used to identity a user [83]. Paruchuri and Chellappan [86] proposed an

architecture using the sensor data from the smartphone to generate the context and

identify the user to the outside world. Various context information such as location,

phone call history, Web browser history etc. are used to identify the user. However,

their architecture is built and ran on a mobile device, thus it cannot work in a

collaborated environment. Kim et al. [87] presented a context-aware platform for

user authentication in Cloud computing. There are three components in the

proposed platform: agent, middleware and Cloud service providers. The agent

works on smart devices. It collects the context information, stores those

information as a user profile. The profile also contains user's preferences. Thus, a

user can initially control the data and specify the object that he wants to reveal his

personal information. Middleware is responsible for connecting the agent and

Cloud service providers. The proposed platform protects users' identity when

authentication with a Cloud service provider, but it does not protect user's online

data.

Biometric authentication methods are also proposed as solutions for mobile

environment. Witte et al.[88] proposed a context-aware mobile biometric

authentication system based on support vector machines. They combine context

information from the embedded sensors on a mobile device and biometric

information together to achieve a higher level security. For instance, a dark

environment may reduce the reliability of face recognition. Thus, context-aware

57
 Chapter 3. Identity Management Systems

information is employed to improve the performance of the biometric systems.

However, their work did not help the user to protect online personal data.

Since the context information is collected from the mobile devices based on the

users' behavior, user-centric model is investigated to solve the security and privacy

issues for mobile users. The user-centric model enables user control their own data

and activities with strong privacy and security protection [89]. Cloud resources

should be collected and allocated according to mobile applications and customized

for each individual user. Bogeret al.[90] proposed a user-centric management

approach using smartcard. In their approach, users' attributes are stored in the

smartcard, thus, a user can control the release of his attributes. However, a user has

to carry the smartcard when he needs to identify himself which cause

inconvenience and increase the possibility of identity theft if the smartcard is lost.

Seo et al.[91] proposed an architecture for context management over pervasive

networks. They set up a centralized context manager which is able to understanding

complex interactions among different contextual entities. Thus, the system can

process and deal with different contexts obtained from different domains around a

user. Their work strengthens the user-centric model by managing contexts from

different providers for the user. Authors in [92] presented an approach empowering

end-users to tailor their mobile applications according to their privacy needs. They

implemented a mobile application prototype, which can collaborate with different

mobile Cloud applications. However, their evaluation results are based on specific

application they developed. The compatibility with existing mobile Cloud apps is

unsolved.

58
 Chapter 3. Identity Management Systems

Based on the research work on context-aware solution and user-centric model,

researchers investigated a new approach to identify the user based on the user

behaviors. The user's context data on a smartphone such as call history, location

records, app usage etc. indicates the user's normal activities. By profiling such

activities, frauds can be detected [93].Lee and Song proposed a model to analyze

user behavior of privacy management on online social network [94].They first

studied the core issues of privacy and then tested conventional theories for the

context in online social network. They contributed to describe the privacy issues

for online social network applications and their proposed model can serve as a

reference for system designers to understand managing privacy on online social

networks. Their research was an ongoing project, and did not consider the privacy

issues for mobile users. Paraskevopoulos et al.[95] developed methods for

enabling the extraction and characterization of normal behavior patterns. By

studying call activity and mobility patterns, they classified the behaviors that

exhibit similar characteristics. The data recorded by cell-towers were also used to

analyze the mobility of mobile phone users. Their experimental results showed

that the proposed methods can be used to predict events and actions that are

possible to happen if some specific circumstances exist. Their research moves one

step forward for detecting users' behaviors. Such evidences can be used to identify

a mobile user. Lathiaet al.[96] presented a platform which is called UBhave, the

UBhave is a project that aims to investigate the power and challenge of using

mobile phones for digital behavior change interventions. Data is collected from

smart mobile devices and user's online social networks activity. By developing

tools to visualize the sequential use of the various interventions over time, the

59
 Chapter 3. Identity Management Systems

overview of user's behaviors can be provided. The analysis of large data set from

the user, the system can be used to promote positive behavior changes. However,

the platform still has issues for the energy constrains at the smart mobile devices.

Furthermore, the privacy concerns when collection data from a mobile user is still

unsolved.

60
3.8 Conclusion

The .NET Passport, Liberty Alliance, OpenID, and Higgins identity management

systems provide the SSO architecture for users‟ convenience when accessing

multiple Web sites in a Web browser. However, such identity management systems

do not have a proper client in mobile platforms which is not suitable for mobile

users. To a certain degree of protecting user‟s online privacy, OAuth is a flexible

and powerful privacy protection scheme that users have the ability to choose the

recipient of his personal data. But it cannot protect the data from abuse by the

hosting service provider. The requirements of mobile Cloud user should not only

focus on identity management, but also on the privacy protection.

Based on the study of the above identity management system and recent

research works, in mobile Cloud environment, users should establish an online

storage which stores all his personal data. A privacy-preserving mechanism should

be deployed to protect users‟ Personally Identifiable Information (PII). The existing

identity management systems can protect user‟s identity from potential attacks, but

naturally they do not consider the privacy protection and thus lack mechanisms to

protect users‟ data. It is necessary to setup an access control system to restrict

access on user‟s online data and using novel cryptographic technologies to satisfy

the requirements in mobile Cloud environment. The behaviour-profiling techniques

can be explored to improve the identification of a mobile user to a more

fine-grained way.

61
4 Access Control Technologies

The exponential growth of Internet usage let user‟s personal data to be

leaked to various unknown service providers. Users do not have control over the

data collected by the service providers and this may eventually breach the user

privacy. Responding to this requirement raises the emerging need of solutions

supporting proper information security governance, allowing users managing their

information to enforce restrictions on information acquisition as well as its

processing and secondary use. Hence, it is necessary to set up a data access control

mechanism so that users‟ data can be protected from being abused. There are three

main solutions that can be explored to protect the privacy of the user's data: (1)

privacy-preserving techniques to protect user data, (2) deploying access control

technique to control usage of user data and (3) storing data in the encrypted form.

This chapter reviews the current main access control models, privacy protection

languages, and encryption schemes, and then discusses the possible solutions to

address the related issues.

This chapter is organized as follows: Section 4.1 discusses background

information about privacy-protection. Section 4.2 reviews the access control

models and the privacy-preserving languages are examined in Section 4.3. Section

4.4reviews the attribute-based encryption schemes and discusses the challenges in

existing approaches. Section 4.5 concludes the chapter.

62
 Chapter 4. Privacy and Privacy-Preserving Technologies

4.1 Access Control Models

Most of the time, users want to safeguard the information that may be harmful or

embarrassing. Thus, the notion of privacy and the notion of control fit together[97].

Access control is a suitable approach for privacy protection in electronic

information systems. Access control defines the selective restriction of access to a

place or other resources. In the information engineering, the access may mean

certain operation (such as Read, Write, Share and Delete) on the resources, or

consuming services. Permission to access a resource is called authorization. With

the years of research carried out over the past decades, several access control

models have been proposed and applied in the industrial applications.

In any access control model, the entities that can perform actions in the system

are called subjects, while the entities which are representing the resources to which

access may be controlled are called objects. In a computer information system, both

subjects and objects should be considered as software entities rather than human

users. Human users may only perform actions on the system via the software

entities which are controlled by them. In most of the current information systems,

access is approved based on successful authentications. The system will make a

decision to grant or reject an access request from an already authenticated subject.

The models that are used by the existing systems are classified into two types,

the capability based model and the access control lists (ACL) based model[98]. In a

capability based model, the subjects will be granted access to any objects based on

an unforgettable reference or capability to the objects. For instances, the possession

of your car key grants you access to your car. Such a capability will be transmitted

63
 Chapter 4. Privacy and Privacy-Preserving Technologies

to another party in order to convey the access. The Plessey System 250 was built in

1970 [99]. It is the first operational computer system to implementation capability

based model. The mechanism is different in the ACL based model. The system

granting access to a subject depends on whether the subject‟s identity is on the

permission list which is associated with the objects. For example, the security staff

of a high-level conference will check a guest‟s ID to see if his/her name is on the

invitation list. Access is conveyed by editing the list. For computer networking, an

access control list refers to rules that are applied to port numbers that are available

on a host, each with a list of hosts that are permitted to use the services.

The capability based model and the ACL based model have the features to

grant access to group members. Such as the doctors in the same department will

have the same access privileges. In a higher level of control, users are assigned with

a user ID. All the processes started by the same user ID have the privileges

corresponding to the user ID. This level of control is still not fine-grained enough.

A doctor has the access to modify any patient‟s records in a department. However,

to a history of records, the system should define the access policy so that they won‟t

be easily tampered. There should be a more fine-grained access control scheme to

achieve a more flexible privacy protection mechanism. In the following subsections,

four main access control models which are categorized as either discretionary or

non-discretionary are discussed [100].

4.1.1 Discretionary Access Control

The Discretionary Access Control (DAC) [101] criteria was defined as follows:

64
 Chapter 4. Privacy and Privacy-Preserving Technologies

“Means of restricting access to objects based on the identity of subjects and/or

groups to which they belong. The controls are discretionary in the sense that a

subject with certain access permission is capable of passing that permission

(perhaps indirectly) on to any other subject.”

The DAC mechanism restricts access to objects based solely on the identity of

subjects who are trying to access them. The user may also be restricted to a subset

of the possible access types (e.g. read, write, execute) available for the protected

resources. Typically, a particular user or set of users have the authority to distribute

and revoke the access to an object, which means a particular user with certain

access permission can pass the permission on to any other subject without notifying

the administrator or the access control system.

The first general model of DAC was proposed by Lampson in 1972 [102].

With several years of research, the DAC mechanism is widely used in variety of

implementations, especially in the commercial and industrial environments because

of its flexibility. In most of the operating systems, such as Windows, Linux, and

Macintosh are based on DAC model[103]. If a user creates a file in one of these

operating systems, he/she can decide the access privileges he/she wants to give to

other users. Then the operating system will make access decisions based on the

access privileges define by the user. However, the DAC mechanism has drawbacks.

For the issue of distributing permission, the particular user cannot provide real

assurance on the information flow in a system. A user who has the access to read

certain data may pass the permission to another user which are not cognizant by the

data owner [100]. The data owner loses the control of the usage of data after

65
 Chapter 4. Privacy and Privacy-Preserving Technologies

sending it to the user. Most of the proposed designs of DAC [104-107]are not

lightweight so that it cannot efficiently deployed in mobile Cloud environment.

4.1.2 Mandatory Access Control

The Mandatory Access Control (ManAC) is commonly discussed in contrast to

DAC as sometimes it is termed as non-discretionary access control. In ManAC

model, a user will be granted access to a resource only if rules exist that allows the

user to access the resource. Subjects and objects have a set of security attributes.

When a subject requests access to an object, an authorization rule will be enforced

by the system, and examines the provided security attributes. A decision will be

made based on the attributes and rules.

ManAC model was investigated over decades. Thomas [108]presented a

ManAC mechanism for the UNIX file system. In this mechanism, all files have a

set of attributes. File attributes consist of name, owner, owning group, access

permissions and modification times of a file, and etc. The operating system uses

some of these attributes to enforce DAC. DAC allows the owners of files to

determine who can access their files by defining DAC access policies. Rjaibi and

Bird [109] presented a multi-purpose implementation of ManAC in relational

database management systems. Their implementation allows a database

administrator to define labels and to set up a database table such that access to a

row in that table is based upon the label associated with that row and the label

associated with the user accessing that row. ManAC are mostly deployed in

centralized systems such as operating systems, and database systems. Li et al.

66
 Chapter 4. Privacy and Privacy-Preserving Technologies

presented a fine-grained ManAC model for XML documents [110]. They

extended XML document model to include label information and define new rules

to satisfy this extended model. An extra XML file contains security labels and

permitted actions are defined. Each request has to satisfy the requirements in the

extra XML files. Authors in [111] presented a ManAC mechanism for mobile

devices based on virtualization. Their proposed mechanism was deployed inside

the virtual machine monitor. The virtual machine monitor is considered as a

trusted computing base. All requests from each subject to objects are enforced by

the access control mechanism based on policies. Thus, mobile devices were not

required to be involved in the access control process. This model moved a step

forward for access control in mobile environment at that time. However, it is out

of date to support mobile Cloud environment.

In practice, a subject is usually a process or thread; objects are constructs such

as files, directories, shared memory segments, etc. For instance, in a database

system, ManAC can also be applied. The objects are tables, views, procedures, etc

[112] and the subject is the user. The root user (the administrator) defines access

rules for each user and the data consent for user is granted based on the rules.

Typically, a centralized security policy administrator will control the security

policies. Users do not have the ability to override the policy, which is contrast to the

DAC mechanism that a subject has the ability to make access decisions and/or

assign security attributes. In real world, almost all UNIX-like operating systems

uses ManAC as a part of the operating system's kernel[113], such as

FreeBSD[114].

67
 Chapter 4. Privacy and Privacy-Preserving Technologies

4.1.3 Role-Based Access Control

The Role-Based Access Control (RBAC) is an approach to restrict system to

authorized users [115, 116]. It is largely adopted in majority of the enterprises and

organizations with more than 500 employees which require multi-level security

[117].

Within an enterprise or organization, roles are created for various job functions,

e.g. manager, director, researcher etc. In a RBAC model, permissions are

associated with the roles. Users are made members of appropriate roles, and

permissions to certain resources are acquired through the role assignments. The

users do not have the access to the resources but only request permission through

their roles. This greatly simplifies the management of permissions. There are

three primary rules defined for RBAC:

Role Assignment: A subject must be assigned a role in order to exercise

permission.

Role Authorization: A subject‟s active role must be authorized for the subject.

This ensures that users can only use roles for which they are authorized.

Transaction Authorization: A subject can exercise permissions only if the

permissions are authorized for the subject‟s active role. This rule ensures that users

can only exercise permission for which they are authorized.

Roles are commonly combined in a hierarchy. High-level roles subsume

permissions owned by sub-roles. Compared with DAC model, access permissions

are controlled by the system not users. RBAC differs from ManAC that is the

management of permissions. ManAC controls basic operation permission (e.g. read,

68
 Chapter 4. Privacy and Privacy-Preserving Technologies

write) based on a user‟s security attributes. While in RBAC, the system controls

complex operation in a transaction (such as an e-bank transaction), or may be as

simple as read or write. In such a view, the management of permissions in a system

is transferred to and can be considered as the management of roles. since a role in

RBAC can be viewed as a set of permissions. Figure 4.1 depicts a standard RBAC

model and Figure 4.2 shows an example of the RBAC role hierarchy.

Figure 4.1Illustration of the Elements in a Standards RBAC Model

69
 Chapter 4. Privacy and Privacy-Preserving Technologies

Figure 4.2Illustration of a Role Hierarchy

Researchers have done many works to improve the RBAC model. Elisa

Bertino et al.[118] proposed the Temporal-RBAC (TRBAC) model which

addresses temporal issues related to RBAC. Coyne and Weil proposed a model

which is called the International Committee for Information Technology Standards

(INCITs) Cyber Security 1.1. The model implemented the access control mapping

between difference information systems [119]. The work improves the

interoperability cross complex enterprises. Elisa and Catania et al.[120] presented a

spatially aware RBAC model which involves location-based services. Their work

extends the RBAC model to deal with spatial and location-based information.

Sejong Oh and Seog Park [121] proposed a task-RBAC model which was founded

on the concept of classification of tasks. The task model deals with each task

differently according to its class, and supports task level access control and

supervision role hierarchy. Strembeck and Neumann [122] presented an context

70
 Chapter 4. Privacy and Privacy-Preserving Technologies

aware RBAC model. Their approach extended the RBAC services to the

enforcement of context constrain. The new requirements for the RBAC can be

found in [123-126]. Kulkarni and Tripathi proposed an context-aware RBAC

model in pervasive computing [127]. They setup a programming framework to

specify and enforce context information during an execution of a role, such as

dynamically interfaced services, permission executions.

Traditional RBAC models lack the ability to specify fine-grained access

control on individuals in certain roles and on individual objects. This is a big

disadvantage when it comes to collaborative environments since their nature

requires permissions based on object types. It suits a centralized environment which

is different from today‟s distributed mobile environments.

4.1.4 Attribute Based Access Control

In Attribute Based Access Control (ABAC) model, access is granted based on

attributes of the user, not based on the rights of the subject associated with a user

after authentication. Unlike RBAC, the ABAC model can define permissions based

on just about any security relevant characteristics, known as attributes [128].

Compared with ABAC, RBAC is outdated, expensive to implement and unable to

accommodate real-time environmental states as access control parameters. ABAC

is newer, simpler to implement and accommodates real-time environmental states

as access control parameters [129].

ABAC fully encompasses the functionality of RBAC approaches. The ability to

make use of additional resource attributes is suitable for mobile Cloud environment.

Attributes such as location, time, date, IMEI number, OS version of a mobile device

71
 Chapter 4. Privacy and Privacy-Preserving Technologies

can strengthen the security level of data transactions and provide more evidence to

the system to make access decisions. Therefore, ABAC can provide a more flexible

and complex access control solution. Due to the mobility of a mobile user, such

contextual attributes keep changing. Traditional access control model cannot utilize

such attributes in run-time. Different combinations of attributes from a subject, an

object and environment make ABAC provide a more fine-grained way than

RBAC.ABAC is a flexible, scalable, dynamic and fine-grained model, with special

features providing great convenience to distributed environments. Moreover, DAC,

ManAC and RBAC are initially designed for access control management within a

single domain, which is not suitable for mobile Cloud environment because of the

collaboration among different mobile Cloud service providers. ABAC is

considered as a suitable access control model for mobile Cloud environments.

Attributes are the set of properties that may be associated with a given entity.

An entity may be a subject, resource or environment that is considered related to the

interaction between a user and an application. Authorization decisions are based on

the related attributes, which are established by digitally signed credentials through

which credential issuers assert their judgments about the attributes of entities.

Because these digital credentials are signed, they can serve to introduce strangers to

one another without on-line contract with attribute authorities [130]. Three types of

attributes are considered for the access control purpose.

Subject Attributes: A subject is an entity which requests access on an object.

The attributes associated with a subject define the identity and characteristics of the

subjects. Such as a subject‟s identifier, name, organization, job, etc.

72
 Chapter 4. Privacy and Privacy-Preserving Technologies

Resource Attributes: A resource is an entity that is requested by a subject. The

attributes of resources can be leveraged to make access control decision with

subjects‟ attributes. The online resources‟ metadata can be considered as relevant

attributes for making decisions.

Environment Attributes: The environment attributes are used for describing

the related operational, technical, and even situational environment or context in

which the access request taken place. Such as time, location, data, operating system

(OS) .etc. The environment attributes are not associated with a subject or an object.

Researchers [128, 130,131] have addressed the issues for deploying ABAC in

the Web service architecture. In [84], Bo Lang and Ian Foster et al. presented an

ABAC model to address privacy issues for Grid Computing. They also provided a

toolkit that is called Globus Toolkit release 4 to implement the framework. Bobba

proposed an attribute-based message system using ABAC to realize a real-time

message system [132]. A location aware ABAC system is presented by Isabel et al.

to extend ABAC with dynamic authorizations. Lanjing Wang and Baoyi Wang

proposed an ABAC model for Web services in multi-domain environment [133].

They presented meta-attribute and meta-policy to describe the attributes and polices

in local domain to collaborate with other domains. ABAC is an ideal access control

model for mobile environments, and several solutions have been proposed to

address the privacy issues in mobile environment [134-137].

73
 Chapter 4. Privacy and Privacy-Preserving Technologies

4.2 Privacy-Preserving Languages

In order to address the issues in the privacy practices, different types of languages

are available to represent the human readable policies in more precise and computer

compatible formats [138]. Some languages are designed to help enterprises express

their privacy policies in ways that are more amenable to policy enforcement and

some languages are designed to help users define their privacy preferences.

Privacy policy languages can help with several of the stages involved in

managing privacy policies (e.g. writing, testing, approving, analyzing, and

withdrawing). They were designed to express the privacy controls that both

organizations and users want to express. Most of the privacy policy languages

were designed for specific purposes with specific features and characteristics and

most of the initiatives for designing these languages have occurred in the last fifteen

years. The W3C began the development of the Platform for Privacy Preferences

(P3P) in 1997. CPExchange was developed in 2000 to facilitate

business-to-business communication about privacy policies [139]. Then the

requirements to express internal privacy polices within an organization from

industry made IBM design the Enterprise Privacy Authorization Language (EPAL)

in 2003[140]. And during the same period, the OASIS presented the eXtensible

Access Control Language (XACML) for both privacy and security policies in a

machine readable format [141].

The privacy policy languages are expected to be fairly simple and small.

Therefore they have been designed as light-weight XML markup languages. They

are not expected to perform high-level mathematical operations or complicated

74
 Chapter 4. Privacy and Privacy-Preserving Technologies

flow controls. The next three sections will briefly discuss three main privacy policy

languages.

4.2.1 The Platform for Privacy Preferences

The Platform for Privacy Preferences (P3P) is developed by W3C in 1997. It is

designed to express website privacy policies in machine-readable format[142] and

to grant users more control over their personal information when browsing the Web.

The P3P user agent allow users to automatically be informed of site privacy

practices and to automate decision-making based on the Web sites‟ privacy

practices [143]. Thus, the P3P Preference Exchange Language (APPEL) was

introduced to express user‟s preferences for making automated or semi-automated

decisions regarding the acceptability of machine-readable privacy policies from

P3P enabled sites [144].

A P3P-enabled website will have a set of policies, e.g. stating the uses of

personal information that is gathered from the site visitors. With a P3P-enabled

Web browser, a P3P user can also define a set of policies, e.g. what personal

information can be revealed to the Web sites that they visited. Then when a user

visits a site, P3P will compare what personal information the user is willing to

release, and what personal information the server wants to get. If the two do not

match, P3P will inform the user and ask if he/she is willing to proceed to site and

risk giving up personal information.

Although P3P provides the support for privacy for Web sites, the Electronic

Privacy Information Centre (EPIC) has criticised P3P, referred to the technology as

a “Pretty Poor Policy” [145]. They claim that P3P software is not suitable for

75
 Chapter 4. Privacy and Privacy-Preserving Technologies

average person to understand because of its complexity and difficulty. Many

Internet users are likely to be unable to use the default P3P software. Furthermore,

the P3P framework is not initially designed for supporting Web privacy. In the

mobile Cloud environments, these disadvantages are critical.

4.2.2 Enterprise Privacy Authorization Language

The Enterprise Privacy Authorization Language (EPAL) is used to formalize

privacy authorization for actual enforcement within an intra- or inter- enterprise for

business-to-business privacy control [146]. EPAL services themselves are

exchanging privacy policies and making privacy authorization decisions. In

particular, EPAL concentrates on the privacy authorization by abstracting data

models and user authentication from all deployment details.

EPAL has a rudimentary temporal nature that a request to perform an action

might lead to allow or deny judgement and obligation [147]. In EPAL, an

obligation is usually an action that some agent is required to perform in the future

which controls the purpose of data usage after sending to the requesters. The EPAL

framework does not consider the privacy for the Web application which is the

limitation for mobile Cloud environments.

4.2.3 Extensible Access Control Markup Language

The Extensible Access Control Markup Language (XACML) was formed by the

OASIS standards consortium. XACML defines a privacy policy language using

the attributes of requestors, resources, and environment. It provides standard XML

schema for expressing policies, rule obligations and conditions. Additionally, it

specifies a request/response protocol for sending a request and having it approved.

76
 Chapter 4. Privacy and Privacy-Preserving Technologies

XACML is primarily an attribute-based access control system, where attributes

associated with an entity will be embedded into a XACML access request, while the

access decision is made depending on whether the access requester satisfies the

XACML policy.

The XACML access control framework contains four key main components:

the Policy Administration Point (PAP), the Policy Decision Point (PDP), the Policy

Enforcement Point (PEP), and the Policy Information Point (PIP).

 PAP: The system entity that creates a policy or policy set

 PDP: The component that evaluates the applicable policy and

renders an authorization decision.

 PEP: The system entity that performs access control. It issues

decision request to the PDP and enforces the access decision

received from the PDP.

 PIP: The PIP is where the entire attribute are stored.

Additionally, in most of the systems, a context handler is required to convert

access requests into the native request format of the XACML canonical form and

convert the authorization decisions in the XACML canonical form to the native

response format. Any required obligations should be performed by PEP with the

enforcement of an authorization decision [148].

Figure 4.3 depicts the data flow of the standard XACML frameworks. The

model operates based on the following steps.

77
 Chapter 4. Privacy and Privacy-Preserving Technologies

Figure 4.3 Data Flow of a Standard XACML Framework

(1) PAPs prepare and write the policies and policy sets to be available to

the PDP. These policies and policy sets represent the complete policy

for a specific target.

(2) The access requester sends a request to the PEP.

78
 Chapter 4. Privacy and Privacy-Preserving Technologies

(3) The PEP sends the request for access to the context handler, in its

original request format, including the attributes from all the

stakeholders, such as subjects, resource, action and environment.

(4) The context handler constructs an XACML request context and sends

it to the PDP.

(5) If additional attributes are required, the PDP will request them from

the context handler.

(6) If the additional attributes request received from the PDP, the context

handler requests them from a PIP.

(7) The PIP obtains the requested additional attributes from subjects,

resource, action and environment.

(8) The PIP returns the requested attributes to the context handler.

(9) The context handler can decide whether to include the resource in the

context.

(10) The PDP evaluate the policy after receiving the requested attributes

and resource (optionally) from the context handler.

(11) The PDP sends the response context and access decision to the

context handler.

(12) The context handler forwards the response context to the PEP after

converting it to its original format.

(13) The PEP fulfils the obligations.

79
 Chapter 4. Privacy and Privacy-Preserving Technologies

(14) If access is permitted, the PEP permits the access; otherwise, it denies

access.

Based on the flow, the PDP has no control over the enforcement of the policy

decision. This model for access control depends on having every request for

protected resources go through PEP, which isresponsible for all enforcements. The

implementation of PEP is usually application- or platform-specific, as it is usually

built into the application or the platform on which the application is built.

Any policy must be stated in terms of a specific vocabulary of terms. In

XACML, vocabulary terms are called attributes, and are defined by the particular

applications or domains that use policies – they are domain-specific. The PDP does

not need to understand the domain-specific meaning of each attribute used in a

policy; it needs to understand only how to determine whether the values supplied

for each attribute in an authorization decision request satisfy the conditions for

access specified resource in the policy. In order to do this, the PDP needs to know

only a unique identifier for the Attribute and the generic data type of the values for

that Attribute. It is the responsibility of each application or domain that will be

using policies to define an Attribute to correspond to each of its policy vocabulary

items. This means assigning a unique identifier for the item, the data type of the

values for that item, and the meaning of each value that might be used for that item.

The meaning of the values is not used by the PDP, however, and is significant

only to the application or domain. Thus, the format (syntax) of an Attribute is

domain-independent, even though the meaning (semantics) of the Attribute value is

80
 Chapter 4. Privacy and Privacy-Preserving Technologies

domain-specific. The policy language may itself define some standard Attributes

that are available for use in any policy.

XACML was largely adopted after being presented in 2002 [149]. Compared

with EPAL, the functionality of XACML 2.0 is a superset of EPAL 1.2. The EPAL

differences often result in less functionality than XACML has [150]. XACML is

designed to support centralized or decentralized policy management. It also

supports both general access control and privacy policies, allowing these closely

related policy types to be integrated. XACML is both a more comprehensive access

control policy language than EPAL, and a full- featured privacy policy language.

XACML has already been an OASIS approved standard. With the support of many

organizations, XACML are more open to the public requirements. Whenever an

issue comes, the researchers can react quickly to address them. There is an active

community of users and developers who are continuing to expand improve, and

apply the language [151] which gives a powerful and long life to XACML.

Many scientific research works have been done to address issues in the Internet

privacy protection field. Claudio extended the XACML architecture and modules

for effective credential-based management and privacy support [149]. Anderson

and Devaraj proposed a XACML-based Web services policy constraint language

(WS-Policy Constraints) [152]. Vivyinget al. extended the XACML with policy

negotiation. They explore XACML‟s potential in privacy negotiation and

introduced a policy negotiation point. Their work addressed issues in the access

control negotiation process between Web services [153].

81
 Chapter 4. Privacy and Privacy-Preserving Technologies

Since XACML provides the support for integrating SAML standard [154], the

combination of XACML and SAML gives a more powerful solution for

authentication and authorization. SAML can protect, transport and request

XACML schema instances and other information needed by an XACML

implementation. Ardagna et al. described extensions to the access control industry

standards XACML and SAML to enable privacy-preserving and credential-based

access control [155]. Their extensions allow the requester to learn which attributes

have to be revealed and which conditions must be satisfied, thereby enabling to

leverage the advantages of privacy-preserving technologies such as anonymous

credentials. Hommel presented architecture for the integration of XACML and

SAML, it established an access control framework in Federated Identity

Management systems [156]. Nils and Vladimir describes a location aware access

control model on top of Geographically XACML. It sketches how spatial

separations on duty constraints (both static and dynamic) can be implemented. The

work provides location dependent access control and security enhancing solutions

on mobile devices [157].

XACML can also be used for mobile environments. Qing presented a

XACML-based policy-driven access control for mobile environments [158]. They

introduce a mapping function and assign a unique ID for the mobile users. An

attribute authority takes charge of maintaining the attributes. An access request

from a mobile user in an unknown domain can be involved in the local

authorization process.

XACML is a powerful privacy access control language which can be deployed

in a distributed system. Although in a XACML policy-based system, the message

82
 Chapter 4. Privacy and Privacy-Preserving Technologies

flows must be formatted in the XACML format, it provides huge convenience for

access control that the user can easily define polices to protect data. Such features

can help mobile users to protect their privacy in mobile Cloud environments. In the

next section, another approach to protect users‟ data in mobile Cloud environments

is investigated.

4.3 Attribute Based Encryption

Besides establishing an access control system, another way to protect privacy in

mobile Cloud environment is by storing the data in the encrypted form, thus

minimizing the risk. However, data sharing is one of the most important features of

mobile Cloud computing. Traditional encryption schemes are one-to-one solutions

and cannot efficiently satisfy the user‟s online data sharing requirements.

Encrypting the data in the Cloud alleviates the above problem [159]. Thus, if

the servers which store the data are compromised, the amount of lost information is

limited because the adversary does not have the decryption key to obtain the data in

the plain text. In a traditional public key encryption method, data is encrypted by a

particular individual who has already established a public-private key pair. The

traditional method in mobile Cloud environment limits the ability of users to

selectively share their encrypted data at a fine-grained level, e.g. a patient wants to

grant decryption access to doctors in a specific department of a hospital for his

medical records. This can be done by either encrypting the file in a shared group

public key, or encrypting the file for each doctor. None of these techniques is

83
 Chapter 4. Privacy and Privacy-Preserving Technologies

particularly appealing. There is a requirement for expressing more complex access

policies during the encryption.

Attribute-Based Encryption (ABE) was proposed by Sahai and Waters [160]. It

is a type of public-key encryption scheme in which the key of a user and the

ciphertext are dependent on attributes. In an attribute based encryption system, data

owner encrypts the data by using several attributes. A decryptor can decrypt the

data only if his/her attributes matches the required attributes. Figure 4.4 shows the

architecture of Sahai and Waters‟ ABE scheme.

Figure 4.4Illustration of Attribute-Based Encryption Scheme

Sahai and Waters made some initial steps to solve this problem for expressing

access polices during encryption. In their system, user‟s key and ciphertexts are

84
 Chapter 4. Privacy and Privacy-Preserving Technologies

labelled with sets of descriptive attributes and a particular key can decrypt a

particular ciphertext only if there is a match between the attributes of the ciphertext

and the user‟s key. An authority with access to the master keys will issue different

private keys to users, where a user‟s private key is associated with an access

structure over attributes and reflects the access policy. The decryption algorithm

allows users to decrypt data using their private keys as long as their access policy

specified by their private key permits.

The original ABE construction of Sahai and Waters allows, the authority to

issue the private key based on threshold access policy. In order to decrypt the data,

at least k attributes should be overlapped between a ciphertext and a private key.

This primitive feature limits the expressibility of the access policy. For instance, the

attribute authority monitors and manages n number of attributes in total where 1 ≤ k

≤ n. The data owner encrypts the file with k number of attributes enforced. When a

user requests a decryption key for the file, the attribute authority issues a private

key for the user with dk number of attributes. Only if k ≤ dk, the user can decrypt the

requested file. Sahai and Water‟s model cannot define more complex access

policies, two main extensions of original ABE scheme were proposed afterwards:

Key-Policy ABE (KP-ABE) and Ciphertext-Policy ABE (CP-ABE) to address the

drawbacks of Sahai and Water‟s scheme and provide more functionality.

85
 Chapter 4. Privacy and Privacy-Preserving Technologies

4.3.1 Key-Policy Attribute-Based Encryption

In 2007, Goyal and Pandey [159] developed a richer type of ABE cryptosystem and

demonstrated its applications. In their system, each ciphertext is labelled by the

encryptor with a set of descriptive attributes. Each private key is associated with an

access structure that specifies which type of ciphertexts the key can decrypt. This

method is called as Key-Policy Attribute-Based Encryption (KP-ABE), since the

access structure is specified in the private key while the ciphertexts are simply

labelled with a set of descriptive attributes. Their method supports fine-grained

access control to facilitate granting differential access rights to a set of users and

allow flexibility in specifying access rights of individual users. The data is stored on

the server in an encrypted form; different users are allowed to decrypt the data

based on the access policy.

In the ABE scheme, the access policy scheme is based on the secret-sharing

scheme (SSS). In 1979 Shamir [161]proposed a construction methodology for

secure-sharing scheme in 1979. In his scheme, the access structure is a threshold

gate. If there are t or more parties take part in within N (1 ≤ 𝑡 ≤ 𝑁) number of

parities, they can reconstruct the secret by using their shares. If there are less than t

number of parties, they cannot get any information about the secret. Benaloh [162]

extended Shamir‟s idea to a tree-access structure. The tree consists of threshold and

the interior nodes consist of AND and OR gates. The leafs consist of different

parties. Any set of parties that satisfy the tree can reconstruct the secret together.

Goyal and Pandey proposed a new secret sharing scheme that each private key is

associated with a tree-access structure where the leafs are associated with attributes.

A Secret-Sharing Scheme (SSS) is used to divide a secret among a number of

86
 Chapter 4. Privacy and Privacy-Preserving Technologies

parties. The information given to a party is called a share of the secret. Every SSS

realizes some access structures that define the sets of parties who should be able to

reconstruct the secret by using their shares.

The KP-ABE scheme can be viewed as an extension of Sahai-Water‟s

conventional ABE framework with the embedding of a secret-sharing scheme in

the private key. The authority can specify a more general secret sharing scheme for

monotonic access trees. They also suggested a new ABE scheme which is discussed

in the next section. Figure 4.5 shows the work flow of KP-ABE scheme.

Figure 4.5Key-Policy Attribute-Based Encryption Scheme

4.3.2 Ciphertext-Policy Attribute-Based Encryption

Bethencourt et al.[163] presented a new ABE scheme which is called

Ciphertext-Policy Attribute-Based Encryption (CP-ABE). In CP-ABE scheme,

user‟s private key will be associated with an arbitrary number of attributes

expressed as strings. When a data owner encrypts a message, he/she specify an

associated access structure over attributes. A user will only be able to decrypt the

87
 Chapter 4. Privacy and Privacy-Preserving Technologies

ciphertext if that user‟s attributes pass through the ciphertext‟s access structure.

The access structures are described by a monotonic access tress, which means the

nodes of the access structure are composed of threshold values and the leaves

describe attributes. They suggest the AND gates can be constructed as n-of-n

threshold gates and OR gates as 1-of-n threshold gates. Figure 4.6 depicts a

working setup of the CP-ABE scheme.

Access Structure A’

Attribute Set T

Private key with User Decryption if A’(T’) = 1

Attribute Set T’
Ciphertext

Figure 4.6Ciphertext-Policy Attribute-Based Encryption Scheme

Traditionally, access control is enforced by employing a trusted server to store

data locally. The server is entrusted as a reference monitor that checks a user‟s

credentials before allowing him to access data. Most existing public key encryption

methods allow a party to encrypt data to a particular user, but are unable to

efficiently handle more expressive types of encrypted access control scenarios.

Both the KP-ABE and CP-ABE schemes can efficiently address this problem. A

88
 Chapter 4. Privacy and Privacy-Preserving Technologies

data owner can exert control over who has access to the data he/she encrypts. All

the users who satisfy the requirements are able to decrypt the ciphertext.

There is only one authority that issues the private keys and manages all the

attributes in both the KP-ABE scheme and CP-ABE scheme making this scheme

highly secure. The authority is normally named as central authority or global

authority. It verifies all the attributes or credentials for each user in the system and

issues private keys to the user. Therefore, it is powerful enough to get all the

messages and communications of the user with or without permissions. Users‟

privacy is at risk if the central authority is compromised. This is the major

drawback of the CP-ABE and KP-ABE scheme. Furthermore, since the central

authority issues private keys for all the users, if the system is built as a large and

global scale system, the authority will become a common bottleneck. If the system

spreads the central authority‟s key over several machines to alleviate performance

pressures, it will also raise the risk of key exposure [164].

4.3.3 Multi-Authority Attribute-Based Encryption

Researchers proposed Multi-Authority Attribute-Based Encryption (MA-ABE)

scheme to address the issues of a single authority ABE scheme. In reality, different

entities are responsible for maintaining different sets of information, such as Driver

and Vehicle Licensing Agency (DVLA) managing all the UK driver and vehicle

information, National Health Service (NHS) managing UK‟s patients‟ information.

In a single authority scenario, in order to get medical and driver‟s information, there

must be at least one fully trusted authority that monitors all the information. If the

89
 Chapter 4. Privacy and Privacy-Preserving Technologies

fully trusted authority is compromised, then all the sensitive data that stored at

DVLA and NHS is under risk. There is a requirement to construct an ABE model

where more than one authority can operate simultaneously, each handing out secret

keys for a different set of attributes.

Chase [165] firstly proposed an efficient MA-ABE scheme in 2007. In her

scheme, the data owner has the ability to specify for each authority k a set of

attributes monitored by that authority and a number dk so that the message can be

decrypted only by a user who has at least dk number of the given attributes from

every authority. The scheme allows any number of attribute authorities to be

corrupted, and guarantee the security of encryption as long as the required attributes

cannot be obtained exclusively from those authorities and the central authority

remains honest. However the central authority still exists in the scheme but with

less functionality. It is stilled fully trusted but handles no attributes and only issues

the corresponding keys.

For a user of Chase‟s model, he will be assigned with a global identifier (GID).

The GID has the following properties:

 No user can claim another user‟s GID

 All authorities can verify a user‟s GID

Thus, the GID can be an identifying string for which a user has provable

credentials and it seems likely that such information would be presented when users‟

attributes are verified. If two users request the private keys from two attribute

authorities, the GID enables the authorities to distinguish the two requests in order

to prevent collusion.

90
 Chapter 4. Privacy and Privacy-Preserving Technologies

Each user will send his GID to the central authority and receive a

corresponding key. The central authority will not get any information about the

user‟s attributes. The central authority only issues a setup key for the user‟s GID

and holds the master secret of the system. Although the central authority does not

monitor any attributes, it still be able to decrypt any messages with the master

secret. Figure 4.7 illustrates the architecture of Chase and Chow‟s MA-ABE

scheme.

Chase‟s work[165]made a small step forward than the single authority ABE

scheme, the main drawback of CP-ABE and KP-ABE still remains with the

existence of central authority. In 2009, Chase and Chow presented an improved

MA-ABE scheme [166] based on Chase‟s previous work. In the latest work, the

central authority was removed and it provides secure privacy-preserving techniques

and better security. In the improved framework, GID is still used as the identifier of

users with same properties discussed before. Each pair of attribute authorities

would share a secret key. For each user, he/she will request secret keys from each

authority and combine them to get the decryption key. The authorities are not

allowed to communicate with each other to prevent collusion attacks. Chase and

Chow used anonymous techniques to hide the identity of users which means the

GID is now completely hidden to the attribute authorities. The authorities cannot

track user‟s attributes from the requests to identify the user and then combine their

keys together to get decryption keys of the user. It achieves the pseurandomity,

anonymity and untracebility, therefore, the privacy of the user is protected. They

also provide extensions to the basic framework, such as to support more complex

access structure and variable thresholds gates across authorities.

91
 Chapter 4. Privacy and Privacy-Preserving Technologies

Figure 4.7Illustration of Multi-Authority Attribute-Based Encryption

The improved Chase and Chow‟s framework provides a more practical ABE

scheme. However, the computational cost and communication overheads are high

and complex. The user has to communicate with all the authorities to obtain the

decryption key. In mobile Cloud environment, the processing power may not be

strong enough and the mobile data network may not be reliable for completing all

required communication.

Lewko and Waters [164] proposed a decentralizing ABE scheme. In their

system, any party can become an authority and there is no requirement for any

global coordination other than the creation of an initial set of common reference

parameters. A party can act as an authority simply by creating a public key and

92
 Chapter 4. Privacy and Privacy-Preserving Technologies

issuing private keys to different users that reflect their attributes. Each authority can

function entirely independently and can join or leave freely. The central authority is

removed in the scheme. It is secure against any collusion attacks and it can process

the access policy expressed in any Boolean formula over attributes. However, their

method is constructed on a mathematical model that incurs heavy computational

cost.

Kan Yang and XiaohuaJia [167] designed an access control framework for

multi-authority and proposed an efficient and secure MA-ABE scheme for Cloud

storage. Their work is the first multi-authority scheme based on the CP-ABE

scheme which does not require a central authority. However, they introduce a

certificate authority which is responsible for assigning an unique identity for each

user and an authority identity for each attribute authority. The certificate authority

is also fully trusted.

Li et al. proposed a framework that achieved fine-grained, cryptographically

enforced data access control model based on MA-ABE scheme for the medical

records stored in the Cloud [168]. The patient uploads his encrypted PHI files to the

Cloud storage using KP-ABE scheme; then the doctor or nurse requests the PHI file

based on the attributes received from all the authorities. However, their proposed

framework is based on a single central authority, which may be the bottleneck of

the systems. Akinyele et al. presented a design of self-protecting electronic medical

records using KP-ABE scheme on mobile devices [169]. Their system allows

healthcare organizations to export medical records to locations outside of their trust

boundary. Each encrypted record has its own access policy. They implemented an

iPhone client application for storing and managing medical records offline.

93
 Chapter 4. Privacy and Privacy-Preserving Technologies

However, their work is not suitable for the collaboration environment since it is a

single authority based scheme.

In mobile Cloud environment, real-time context-related attributes can be

captured to provide more complex policies to secure the data access control. These

attributes can be time, current location, date, IMEI number and any other contextual

attributes. Those can be derived from the local context information using the

hardware embedded within smart devices. Enforcing these attributes into the

encrypted data during the encryption can effectively reduce the risk of collusion

attack. Several works are discussed in the following paragraphs. They have

proposed approaches to incorporate such type of attributes.

Liao and Chao [170] proposed a location-based encryption scheme. They

proposed a location-dependent approach to incorporate the coordinates while

encrypting. The user can only decrypt the data when the coordinates acquired for

GPS receiver are matched with the target coordinate. Considering the inaccuracy

and inconsistency of the GPS signal, they involve a toleration distance to increase

the practicality. However, privacy-preserving and data access control were not

considered. A secure, selective group broadcast system in vehicular networks using

dynamic attribute-based encryption is presented by Chen et al. in [171]. Chen et al.

introduced a fading function to CP-ABE scheme in order to allow attribute level

updates. The fading functions allow treating attributes in accordance with their rate

of change so that attributes can be updated independently from the entire private

keys. By choosing a proper fading rate, the balance of trade-offs between security

and efficiency can be controlled to deliver a better broadcasting service or a better

security services. Their work finds a solution for using context-related attributes

94
 Chapter 4. Privacy and Privacy-Preserving Technologies

under a single domain. Xu and Martin proposed a dynamic user revocation and key

refreshing for ABE in Cloud storage [172]. In Xu and Martin‟s model, the system

can refresh the system keys or remove the access from a user without issuing new

keys to other users or re-encrypting existing ciphertexts. Therefore, the mobile user

can have a more flexible solution to protect their data and privacy in the mobile

Cloud environment.

4.3.4 Challenges

In the ABE scheme, collusion attack is the most important challenge. For example,

Alice is a doctor in hospital A and Bob is a student at City University London.

Sender encrypts a message with two out of three of the following attributes: doctor

of hospital A, student at City University London, and home address in London. If

Bob and Alice combine their attributes then they have two out of three required

attributes, but they should not be able to get the decryption keys to obtain the

content of the message.

Sahai[160] addressed this issue by generating different random sharing of a

secret for each user. So private keys issued for different users cannot be combined.

In Chase‟s original MA-ABE scheme [165], GID is introduced as the identifier of

each user and can be distinguished by each authority to prove the possession of

attributes. In MA-ABE, the scenario is different from the single authority.

Assuming that there are two authorities, AA1and AA2. Alice request keys for

attribute set S1 from AA1and S2 from AA2. Then Bob also requests keys for

attributes S1 for AA1 and S2 from AA2. If the authorities do not exchange

information about the two requests and Alice and Bob are identified by nothing

95
 Chapter 4. Privacy and Privacy-Preserving Technologies

beyond their attributes. So from the authorities‟ point of view, the two scenarios

must be identical. By using the GID, the authorities can distinguish these two

scenarios in order to prevent collusion attacks.

However, using the GID may cause the system to issue the decryption keys

based on the identifier and not based on the attributes. So Chase still uses the central

authority to issue a setup key for the user‟s GID. The central authority will not

monitor any attributes; user only sends the GID to it and receives the key. Each

authority computes his own independent secret by giving away only the GID. With

the key obtained from the central authority, the user can always combine the results

to obtain the secret. The secret given by each authority is a pseudorandom secret

and only combined with the key from the central authority.

In Chase and Chow‟s improved MA-ABE scheme [166], without the adding of

additional key of the central authority, each attribute authority generates a set of

secret while issuing keys. When user receives these keys and reconstructs the secret,

the computation will use these secrets to get the summation which is equal to 1 or -1.

The innovative mathematical design successfully addressed these issues. MA-ABE

is an ideal solution for data security in the mobile Cloud environment, since it

combines the techniques of access control and cryptographic technologies..

96
4.4 Conclusion

A large amount of personal data is generated by mobile-based services; these

user-generated data must be properly protected from unauthorized usage. Advances

in technology and the rapid use of mobile Cloud computing systems created a

necessity for protecting context sensitive information transferred through the

system. The access control is not only for defining who is able to take actions on

resources, but also to control the purpose for which the resources were accessed. In

mobile Cloud environments, there are real-time contextual attributes which can be

derived from the context information. From this aspect, the access control model

for a mobile use case should be able to capture not only the static attributes of

related stakeholders, but also dynamic attributes to improve the security level.

The DAC, ManAC and RBAC are designed for a traditional

privacy-preserving scenario. They can still continue to contribute to suitable

environments. In today‟s mobile Cloud environment, user‟s mobile devices can be

considered as a node on the whole network. Those smart devices are always

provided with Internet access through Wi-Fi or 3/4G data networks. They are

powerful enough to execute any Web-based or Cloud-based services. Security and

privacy-preserving framework is required to secure the communication between

two parties and also to protect the collected personally identifiable information in

the connected digital society. ABAC is a flexible, scalable, and fine-grained access

control model which has shown enormous potential in the last couple of years to

enhance the service provisioning from smart mobile handsets. Considering the

97
 Chapter 4. Privacy and Privacy-Preserving Technologies

nature of attribute-based authorization, XACML is an ideal privacy policy

language to work with ABAC. The combination of XACML and ABAC can

address privacy-related issues in the mobile Cloud environment.

The MA-ABE scheme is a suitable solution for addressing issues of both

security and privacy aspects in mobile Cloud environment. This chapter examined

the privacy protection technologies of attribute-based encryption schemes for

encryption and access control of personal data. Existing privacy protection

techniques are not designed for mobile environments. The work proposed in this

thesis will combine state of the art existing access control and privacy preserving

techniques to provide a novel framework through which the personal data can be

protected in a mobile Cloud environment.

98
5 User-Centric Attribute-Based Access Control Model
Using XACML

The exponential growth of mobile applications let users to leave a substantial

amount of online data via Cloud-based services such as Google services inthe

Internet. As discussed in Chapter 3, leaving the data (personal) in online Cloud

storage raises privacy issues. There is a need for user-centric approach whereby

users can control the usage of their personal data in order to mitigate privacy

breaches. Therefore, in today‟s mobile Cloud environment, privacy protections are

important and technologies for protecting privacy are considered as vital features in

mobile Cloud-based services. Hence, in this Chapter the author proposed a general,

flexible approach to support the user-centric access control model based on

XACML, where user controls the privacy of personal data in the mobile Cloud

environment.

This chapter provides an implementation of a user-centric access control model

using XACML. Section 5.1 describes the architecture of the model and Section 5.2

gives details of policy evaluation process. Section 5.3 gives the security evaluation

of the proposed model followed by a case study. The proof of concept and

implementation details are given in Section 5.4. The conclusion is given the

chapter in Section 5.5.

99
 Chapter 5. User-Centric Attribute-Based Access Control Model Using XACML

5.1 Architecture of Policy-Based User-Centric Approach

In a mobile Cloud environment, the following contextual attributes can be exploited

to secure the personal data: location, time, date, app usage etc. These attributes can

be exploited to authenticate the user who requests permission for data access. In the

proposed approach, authorization is a mutual process between a user and a service

provider (SP). To allow a SP to access the personal data, the user authorizes the

SP by pre-defined privacy access policies while the user is also challenged by the

system in order to verify whether he/she is the actual user who is requesting

services from the SP. An Attribute Authority (AA) which stores all the users‟ and

SPs‟ attributes for authorization is introduced. The AA is a trusted party by both the

users and SPs. By defining personal privacy access policies, users can control the

disclosure of sensitive information which is stored online (e.g. Cloud based

storage). Thus, the proposed model enables the mobile users to authorize the SP

with the participation of AA. By using the available context-related attributes, the

system also verifies the user to ensure it is the actual person who is requesting the

services. Compared with the current identity management systems, the proposed

model focuses on using context-related attributes to verify a mobile Cloud user at

run-time which is more flexible and secure in mobile Cloud environment, and

provides data access control for the mobile Cloud user to enforce access control

on his/her online data.

100
 Chapter 5. User-Centric Attribute-Based Access Control Model Using XACML

In the proposed model, the users define XACML policies in order to control

the access to their personal online data. XACML is a privacy access policy

language which is a dialect for defining attribute-based access control policies

[173]. The proposed user-centric approach has five main actors: the user, the SP,

Policy Evaluation Component (PEC), the AA, and Identity Provider (IdP) as shown

in Figure 5.1.

In the proposed model, AA is considered as a central authority. It maintains the

attributes of the users and SPs. SP has to register with AA in order to provide a

service to the mobile Cloud user. AA assigns a unique Service Provider Identity

(SPId) to the SP and this identity is stored at the AA. Any attributes related to the

SP is also stored at the AA. In most of the current identity management systems, an

identity provider authenticates the users and SPs[174]. It is also introduced in the

proposed model, and it is not only responsible for authentication, but also connects

the users, SPs, and the PEC. In order to restrict access control on users' online data,

PEC is introduced to enforce access control. PEC is modified based on the standard

XACML structure. It extends the existing XACML standard [151]with real-time

authentication for the user. By cooperating with AA, real-time attributes such as

location, time are used to verify the user. Therefore, the proposed model enables

users to protect their online data by personal preferences in real-time.

Details of each actor are briefly explained below:

 User: A user is a person with a mobile device and the mobile device

is connected to the Internet via the mobile data network or Wi-Fi.

101
 Chapter 5. User-Centric Attribute-Based Access Control Model Using XACML

 Attribute Authority: The AA is the trusted party that stores

attributes representing the SPs and the users.

 Policy Evaluation Component: PEC is the core component that

formats all the requests into XACML requests, evaluates the access

requests and performs the access decisions.

 Identity Provider: IdP is responsible for authentication of the user

and forwarding the access requests to the PEC.

 Service Provider: A SP provides online resources (such as Cloud

services or online data) to the user and requests user‟s data based on

the policies.

Figure 5.1 Framework of the proposed User-Centric Policy-Based Access

Control Model Using XACML

102
 Chapter 5. User-Centric Attribute-Based Access Control Model Using XACML

The user‟s personally identifiable information was defined as PII in the

previous chapter. In order to authorize the user, PII is used as key attributes for the

user authorization process. The U.S. Office of Management and Budget suggests

several examples of attributes related to the PII [175] as given below:

 Name: Full name, maiden name, mother‟s name, or alias.

 Personal Identification number: Social security number (SSN),

passport number, driving licence number, taxpayer identification

number, or financial account or credit card number.

 Address Information: Street address, home address, IP address, or

email address.

 Personal Characteristics: Photographic image (especially face or

other identifying characteristic), fingerprints, handwriting, or other

biometric data.

 Information about an individual that is linked or linkable to one of

above: date of birth, place of birth, race, religion, weight,

geographical indicators, employment information, medical

information, education information, and financial information.

In some countries, people are assigned with a unique number which can be

used for identity verification. In the UK, driving licence number and National

Insurance (NI) number can be considered as unique numbers. As studied in Chapter

4, the Global Identity (GID) is considered as a key attribute in the proposed model

103
 Chapter 5. User-Centric Attribute-Based Access Control Model Using XACML

which stands as an identity of a user. The user is registered using the GID and other

attributes. Assuming that the GID is a secret and only the user knows the exact

value of the GID.

The IdP and the PEC have an existing trust relationship. The user is assumed to

request data and services from the SP. The IdP can also be a federated identity

provider whereby the user can use the same identity to access any other partner sites.

The user needs to register with IdP before requesting online resource from the SP.

PEC is an online access control service, which enable users to define access control

policies for their online data. PEC can be considered as a private Cloud that

implements access control services on behalf of the user. AA is trusted by the user,

SPs, and PEC.

5.1.1 System Initialization

In the proposed access control model, there are several steps that needs to be

executed by both the user and the SPs in order to initialize the system.

Initially, a user needs to finish the registration process. The registration process

can be done online. The registration process includes the following steps:

(1) Register with the IdP

(2) Register with the AA

(3) Define privacy access policies for his/her attributes at the PEC

After the registration, the user‟s details such as username and password will be

stored in the IdP, the data which is related to users‟ privacy is stored at the AA. For

104
 Chapter 5. User-Centric Attribute-Based Access Control Model Using XACML

instance, the NI number and driving licence number are stored at the AA. With the

pre-defined XACML policy, the PEC can evaluate any access request on user‟s PII.

For a SP, the registration process is explained below:

(1) Register with the AA and assign a SPId

(2) Register with the IdP

After the registration, the SP‟s attributes such as SPId, name, address etc. are

stored at the AA. SPId is considered as an identifier of the SP.

5.1.2 Design of the Model

The main goal of proposed model is to enable mobile Cloud usersto define access

control polices on their online PII. Thus, data requesters who satisfy the access

policy can be granted access on these sensitive personal online data. Assuming all

the actors of the proposed model and any services or data provided by the system

must be available online. The workflow of the proposed model is depicted by the

stages described below:

Stage one: for a registered user, the first step is to login into the system via IdP.

After receiving the request, IdP starts the authentication process. If the user is

successfully authenticated, he can then request any data or services from a SP. Once

received the resource request, the SP would ask several attributes of the user in

order to deliver the resource. In a mobile environment, the user should have full

control over his online personal information. Thus, in order to protect user‟s

privacy, an access control mechanism which restricts the access on user‟s personal

105
 Chapter 5. User-Centric Attribute-Based Access Control Model Using XACML

data is proposed. The SP must satisfy the privacy access policy defined by the user

in order to gain the PII data. The policies stored at the PEC are used to control the

access on user‟s data and the policy evaluation is carried out by the PEC.

Stage two: after requesting the data or services, the IdP generates a local

security context based on the request. The security context contains information of

the user, SP and the requested resources. The security context will be forwarded to

the PEC by IdP.

Stage three: after receiving the security context, PEC extracts information

about the user, SP and the requested resource from the security context. PEC

executes the following three steps:

Step 1: Retrieve the information of the SP, query the AA for the

relevant attributes of the SP in order to check the authenticity of the SP.

Step 2: Retrieve the information of the user, query the AA for the

relevant attributes of the user to check the authenticity of the user.

Step 3: Evaluate the request based on the policies stored in the Policy

Repository.

In order to verify the user in Step two in Stage three, the PEC needs to generate

a challenge message and send it to the user. User‟s mobile device generates a

response message based on user‟s inputs. Then PEC starts the policy evaluation if

the verifications of the SP and user are successful. If the result of policy evaluation

in Step 3 is successful, then PEC sends the result to the user and SP via IdP.

From the three stages, the proposed model support SSO which is an important

feature in today's identity management systems. The use of real-time attributes

106
 Chapter 5. User-Centric Attribute-Based Access Control Model Using XACML

enables the proposed model to dynamically authenticate the user, which cannot be

provided by the current identity management system such as OpenID, Liberty.

Figure 5.2 depicts the workflow of the proposed model. The message flow is

described below.

Figure 5.2Illustration of the Message Flow of Proposed User-Centric

Policy-Based Access Control Model

(1)The user sends login requests to the IdP using the mobile device.

107
Chapter 5. User-Centric Attribute-Based Access Control Model Using XACML

(2)The IdP sends challenge message to the user for authentication.

(3)After the user inputs username and password, the mobile device

generates a response message.

(4)The IdP authenticates the user based on the response message.

(5)The user sends a request to SP to access data or services.

(6)The SP forwards the attribute of itself and the attributes of the

requested resource to IdP

(7)User sends his attributes to the IdP

(8)The IdP generates a local security context based on the request. The

context contains necessary attributes of the SP, the user and the

requested resource.

(9)The IdP forwards the request to the PEC for further policy

evaluations.

(10)The PEC receives the request and extracts any required information

from the request. Then, the PEC queries the AA to check the validity

of the SP.

(11)The PEC queries the AA for the GID of the user and generates a

challenge request to the user based on the GID.

(12)After receiving the challenge, the user has to answer the challenge.

The mobile device generates a response message which contains the

user‟s answers and sends the response back to PEC.

108
 Chapter 5. User-Centric Attribute-Based Access Control Model Using XACML

(13)If the response is verified, then the PEC starts to evaluate the

request.

(14) If the request is approved, the requestor will be granted permission.

The decision will be sent to the IdP.

(15) The IdP forwards the decision to both user and service provider.

From Figure 5.2, it shows that once the mobile user requested the data or

services from SPs, IdP starts to collect all the necessary information from the user

and SP. The information from the user contains username, or location, time and any

other dynamic real-time attributes. Meanwhile, the information also contains

details of the SPs and the information about the requested resource. Such

information consists of related attributes and the IdP generates a local security

context based on these attributes. Thus, the security context consists of three parts:

(1) Attributes of the user

(2) Attributes of the SP

(3) Attributes of the requested resources

The PEC extracts necessary information from the security context and queries

the AA for any information about SP. If the response from the AA shows that the

SP is a legitimate provider, then the PEC checks whether the user has defined any

access control policy for his personal data.

Based on the requested resource, the SP asks any PII attributes of the user in

order to deliver the resource. During this step, the PEC challenges the user in order

109
 Chapter 5. User-Centric Attribute-Based Access Control Model Using XACML

to authenticate the user‟s identity. Policy evaluation protects user‟s privacy. Such

operations that are carried out at the PEC are the necessary steps for policy

evaluation process of the system. In the next section, more details about the

architecture of PEC and the policy evaluation process will be presented.

5.2 Policy Evaluation Component (PEC)

Policy Evaluation is the core process of a privacy policy-based access control

system. In our model, PEC is the key component of the architecture. It is

responsible for XACML authentication and authorization. After receiving the local

security context, PEC starts to the access evaluation for the access requests.

The PEC consists of the following main functional components:

CH: Context Handler, formats received access requests into XACML format

PEP: Policy Enforcement Point, issues access requests to the PDP and

enforces access decisions

VP: Validating Point, verifies the SP

PDP: Policy Decision Point, evaluates the applicable policy and renders an

authorization decision

RG: Request Generator, generates challenge requests to verify the user

PR: Policy Repository

110
 Chapter 5. User-Centric Attribute-Based Access Control Model Using XACML

Each component performs different functionalities and processes. Let us recall the

local security context generated by IdP in the previous section. Once the PEC

received the security context (see step 8 in Figure 5.2), the PEP extracts the

attributes of the SP and user. Then PEC queries the AA for the attributes of the

mobile user and SP. Figure 5.3shows the architecture of the PEC.

Figure 5.3Architecture ofthe PEC

111
 Chapter 5. User-Centric Attribute-Based Access Control Model Using XACML

Figure 5.4 depicts the authorization process of the model. For instance, if a

mobile user wants to purchase an item from the SP, then the user needs to request

the mobile banking services to complete the payment. Generally, the user is

concerned about the authenticity of the SP. At this time, the user needs to check

whether it is the real SP with whom he/she wants to communicate. The VP

compares the attributes of the SP obtained from AA and the attributes extracted

from the security context received from IdP. If the SP is verified, RG will generate

a challenge message. PEP forwards the challenge message to the user via IdP.

User‟s mobile device displays this challenge request and waiting for the inputs from

the user. This step is to ensure that it is the actual account holder who is requesting

the resource. A response message will be generated using the input and sent to the

PEP. PEP forwards the response to VP. If VP successfully validates the user, CH

will generate a XACML access request and send it to the PDP for policy evaluation.

PDP makes the access decision based on the policies in the policy repository. If the

access decision is “Deny”, then user‟s personal data will not be revealed to the SP.

112
 Chapter 5. User-Centric Attribute-Based Access Control Model Using XACML

Figure 5.4Illustration of Work Flow of Policy Evaluation

The workflow of PEC is briefly depicted below:

(1) IdP generates the security context and sends it the PEC

(2) PEC retrieves attributes from AA

(3) PEC starts to check whether the SP is a legitimate service provider or

not

113
 Chapter 5. User-Centric Attribute-Based Access Control Model Using XACML

(4) A challenge message is generated by PEC in order to verify the user,

the challenge is forwarded by the IdP

(5) User‟s mobile device generates a response message based on user‟s

input and sends it to the PEC via IdP

(6) PEC starts to verify the user based on the response message

(7) If the verification is successful, PEC then start the policy evaluation

process

(8) PEC sends the access decision to IdP

(9) IdP forwards the decision to both the user and the SP

This approach enables user to define privacy access policies to protect their

data. In order to deliver data and services, SP requires user's PII. Current identity

management systems cannot provide flexible access control mechanisms to help a

mobile user restrict access on his/her PII. The proposed model is designed and built

on top of the standard XACML framework. Users can simply define privacy access

polices to control access from different requests on his online PII. The thesis

contributes the access control model for mobile Cloud users by modifying the

standard XACML framework. The access control evaluation is given in next

subsection.

114
 Chapter 5. User-Centric Attribute-Based Access Control Model Using XACML

5.3 Security Evaluation

5.3.1 Protocols on Authentication

5.3.1.1 Protocols on Authentication

Based on the message flow presented in Figure 5.2 and Figure 5.4, the protocols
on authentication can be separated into the following stages.

 User (smart devices) authenticates to the IdP

(1) The user will request access to the SP from the IdP.

(2) IdP challenges the user with username and password.

(3) IdP verifies the user and redirect the user to the SP

(4) The user log into the SP

Figure 5.5Authentication between User and SP

 SP authenticates to the PEC

(1) SP sends attributes to IdP

(2) IdP generates local security context

115
 Chapter 5. User-Centric Attribute-Based Access Control Model Using XACML

(3) IdP sends local security context to PEC

(4) PEC verifies SP

Figure 5.6Authenticaion for SP

 User (smart devices) authenticates the PEC

(1) User sends attributes to IdP

(2) IdP generates local security context

(3) IdP sends local security context to PEC

(4) PEC challenges user using GID

(5) User responses to the challenge

(6) PEC verifies the user

Figure 5.7Authentication for User

116
 Chapter 5. User-Centric Attribute-Based Access Control Model Using XACML

5.3.1.2 Security Parameters

 Local Security Context

Local Security Context is a set of attributes from user, SP and the required
resources. The PEC validates user and SP based these attributes. It is
defined in XML format.

<SecurityContext>

<user>

<username>John</username>

<time>201206121230</time>

<organization>City University London</organization>

</user>

<SP>

<SPId>200010000001</SPId>

<SPName>Citi Bank</SPName>

<time>201206121235</time>

</SP>

</SecurityContext>

5.3.2 Security Analysis

The current identity management systems and existing technologies provide

powerful protection to the data confidentiality and integrity. The proposed model

uses existing cryptographic technologies to secure data communications between

each active role. The main goal of the proposed model, is to utilize the existing

117
 Chapter 5. User-Centric Attribute-Based Access Control Model Using XACML

dynamic attributes in a mobile Cloud environment to protect users' online data

based on pre-defined privacy access policies.

5.3.2.1 Access Evaluation

Let consider the following simple scenario: A customer who lives in the UK is

travelling to Paris. During the travel, he wants to use his mobile device to access the

online service from a SP in Paris and at the same time would like to reject all

requests for online transactions from the UK based ISPs during this period. The

security analyse is given in the following subsections.

Our proposed scheme is secure from spoofing attacks. Use the above scenario as a

typical example and assume the user has defined an access policy that any access

requests from Paris can be approved. The flow is based on Figure 5.2.

User: Login to the IdP in Paris.

Adversary: Registered at the IdP as a service provider. The user can request data or

services from the adversary. Adversary want to collect users' online data without

permission.

(1) User request a service from the adversary

(2) IdP collects attributes from the user and adversary

(3) IdP collects access request on user's data from adversary

(4) IdP generates local security context and sends to PEC

(5) PEC extracts attributes from the local security context

118
 Chapter 5. User-Centric Attribute-Based Access Control Model Using XACML

(6) PEC queries AA to get legitimate attributes for AA

(7) PEC verifies SP based on the attributes received from AA

(8) PEC verifies the user based on the GID

(9) The access request is evaluated by PEC

(10) PEC sends the results to IdP

(11) IdP processes the results

Initially the user is logged in to the IdP. IdP collects the necessary attributes from

both the user and SP. Attributes from user are location, time, username, OS, IMEI

etc. Attributes from the SP are SPId, name, time, IP address, location etc. IdP

generates the local security context, which is described below:

<SecurityContext>
<user>
<username>user</username>
<time>2013007291600</time>
<location>Paris</location>
<OS>iOS 7.0</OS>
</user>
<SP>
<SPId>200010000001</SPId>
<SPName>Thomas Cook</SPName>
<time>2013007291600</time>
<location>London</location>
<IP>0.0.0.0</IP>
</SP>
<acessrequest>
<subject>Thomas Cook</subject>

119
 Chapter 5. User-Centric Attribute-Based Access Control Model Using XACML

<object>
<name>John</name>
<gender>Male</gender>
<...>
......
</...>
</object>
</accessrequest>
</SecurityContext>

PEC extracts necessary attributes from the local security context. In order to verify

the user and SP, it queries the AA for correct attributes to evaluate access request.

Recall that the user has defined access policy on his personal online data. The

adversary has to pass the following two steps.

 PEC verifies SPId based on the AA's records

Adversary's SPId which is sent via the local security context needs to

match the record from the AA.

 PEC evaluate the access request based on the user's pre-define policies.

If the SP passes the first step, PEC evaluates the access request. In this

scenario, the request is: ''Thomas Cook requests John's online PII in UK'.

The local security context proves user's current location is Pairs, thus the

access policy is active. Based on the polices, the access decision is

'FALSE'.

The adversary can gain access to user's online data unless 1) it is a legitimate SP

which has records in AA, and 2) its access requests satisfy user's privacy access

120
 Chapter 5. User-Centric Attribute-Based Access Control Model Using XACML

policy. Therefore, the proposed model provides extra protection on user's online

data.

Following the verification of the AA, PEC verifies the user based on the GID. The

challenge message is displayed on the smart devices and asking user to input the

GID. The consideration for this step is unauthorized use of the smart device. The

'unauthorized use of the smart device' suggests that someone is trying to use the

smart device without the user's permission. For instance, user lost the device in

Pairs, someone got hold of the phone and ordered products on user's behalf. GID is

sensitive information and only known to the user.

5.3.2.2 Security on Communications

Communication security is mainly focusing on the message confidentiality and

integrity in the proposed mobile environment, which is evaluated with respect to

the encryption or signature algorithm and the cryptographic keys.

Roles in the proposed model such as SP, IdP, PEC and AA can be considered as

online services providers．The communications between these entities are normally

carried out using the public key infrastructure. Each entity has a private key and

public key pair. The sender use its own private key to sign message and use the

receiver's public key to encrypt the message. Thus, the receiver uses sender's public

key to verify sender digital signature and decrypts the message using its own

private key.

With the development of mobile platforms, smart devices can provide more secure

and flexible ways to secure communication. Android and iOS supports mobile apps

using the SSL to secure the communications [176]. Hu et al. proposed a 3 factor

121
 Chapter 5. User-Centric Attribute-Based Access Control Model Using XACML

(PIN code, SIM card authentication and facial biometric) authentication method to

protect the message security on Android platform [177].In the proposed model, the

mobile user uses these existing technologies to protect message communications,

however, provides more data access restrictions on his online data usage.

5.3.3 User-Centric Approach

In the proposed model, the IdP authenticates the users. If the user is successfully

authenticated by the IdP, then he can request data or services from the SPs. A

registered user trusts the AA and pre-defines the data access control policy for his

PII data. The user initiates the access request for targeted resources from a SP; the

IdP generates the local security context based on the request. The context consists

of user‟s information (e.g. username), and SP‟s information (e.g. SPId).

In order to deliver the requested resources to the user, the SP requests

information about the user. In an e-health service use case, the SP needs to know the

historical medical information of the user in order to provide medical advice. This

raises a security issue: which part of information can be revealed to the e-health

system? Most of the existing systems lack mechanisms to protect the user‟s PII. In

default, SPs collect any information or allow a partner site to collect the required

information. Hence, user has no control over the data once it leaves his premises.

In our proposed model, more sensitive attributes are stored at the AA. AA has the

ability to control the attributes disclosure. XACML privacy access policies can be

defined by the user at PEC. Thus the user can control the disclosure of his PII from

unauthorized access.

122
 Chapter 5. User-Centric Attribute-Based Access Control Model Using XACML

On receiving the security context from the IdP, the PEC extracts the

information from the security context and then the PEC queries the AA to check the

legality of the SP. Since the SP is registered at the AA with required information,

the AA can validate the SP. This step ensures that the SP is a registered and

legitimate provider. If the SP is verified, then the PEC queries the AA for user‟s

information. In this scenario, the GID is considered as the identifier of the user. The

following two phases are carried out to protect the user‟s privacy:

 SP acquires user‟s attributes to satisfy its own policy in order to

deliver resources to the user.

 PEC receives the request from IdP, evaluates user‟s pre-defined

policies to decide whether to disclose the requested data.

These two phases are crucial for the mutual authorization and solve the

potential authorization negotiation between the user and SP. The pre-defined

policies have the higher priority than the policy defined for the requested resources.

If there is a conflict that the SP is requiring sensitive information in which the

pre-defined policies are forbidden, then the PEC rejects the SP‟s request to protect

user‟s privacy.

5.3.4 Use Case Study

Many hospitals today are providing mobile services to the patients. The e-health

service is a typical user case for mobile Cloud users. For example, a patient wants to

choose a doctor in dental department. In order to serve the patient, the hospital

requires his personal information. He firstly registers with the hospital system as a

patient and uploads the PII and related medical records to the Cloud Storage, which

123
 Chapter 5. User-Centric Attribute-Based Access Control Model Using XACML

can be managed by the hospital system administrators. And then he defines an

access policy that only a doctor in dental department can read his PII and the

medical records.

In this scenario, the patient requesting the services from the hospital can be

considered as the user in the proposed model. The Cloud storage is the AA who

manages attributes for both the users and doctors. There is a central server which

enforces access control and is responsible for the verification of user and SP. The

central server can be considered as the PEC.

After logging into the system, the patient requests for a dental service. A doctor

responds to the request and chooses to treat the patient. The identity provider of the

system receives the requests and collects attributes from the patient and the doctor.

A local security context is created and sent to the central server. The central server

extracts the attributes from the local security contexts. Based on these attributes, it

queries the AA in order to verify the doctor and the patient. The verification method

can be defined by the practical requirements. For instance, Doctor can be verified

by the staff ID, job title, and department. The patient can be challenge not only by

the GID, but also through security questions defined by himself. Once the

verifications are successful, PEC evaluates the access request from the doctor.

Neither a doctor from a department other than the dental department, nor a nurse

from the dental department can gain access to user's PII and medical records. This

information is available to a doctor from the dental department.

Thus, the patient restricts access on his online sensitive data. Privacy and data

security are protected by unauthorized access. The policy-based access control

124
 Chapter 5. User-Centric Attribute-Based Access Control Model Using XACML

provides a simple method to the patient if he wants to disclose personal data to a

specified entity.

5.4 Proof of Concept

5.4.1 Protocol Verification

This section provides formal protocol verification and validation for the proposed

model. Scyther tool [178-180] was developed by CasCremers for security protocol

verification. It is a tool for the automatic verification and falsification of security

protocols. Scyther provides a graphic user interface which incorporates the Scyther

command line tool and python scripting interface.

The syntax of Scyther tool will not be discussed in the thesis, but some basic
knowledge is explained in order to explain the protocol validation process.

send_1(X, Y, data1): role X sends data1 to role Y

recv_1(X, Y, data1): role Y receives data1 from role X

claim_x(X, Secret, data1):role X claims that the event of 'sending data1 to

role Y' is secret.

The five actors in the proposed model is defined as role in the Scyther
validation tool. The message flow sequence in Figure 5.2 is followed and
implemented in the Scyther tool as follows:

protocolProtocolValidation(User, SP, IdP, PEC, AA)

role User {

constuname: Data;

125
Chapter 5. User-Centric Attribute-Based Access Control Model Using XACML

freshattrU: Data;

var challenge: Nonce;

//user uses SSL to send username

send_1(User, IdP, {{uname}pk(IdP)}sk(User));

//user sends real-time attributes

send_2(User, IdP, {{attrU}pk(IdP)}sk(User));

//receives challenge from PEC

recv_7(PEC, User, {{challenge}pk(User)}sk(PEC));

claim_user(User, Secret, uname);

claim_user(User, Secret, attrU);

claim_user(User, Secret, challenge);

role SP {

freshattrSP: Data;

//SP sends attributes to IdP

send_3(SP, IdP, {{attrSP}pk(IdP)}sk(SP));

claim_sp(SP, Secret, attrSP);

roleIdP {

varuname: Data;

126
Chapter 5. User-Centric Attribute-Based Access Control Model Using XACML

varattrU, attrSP: Data;

freshLSConxt: Data;

//IdP sends local security context to PEC

send_4(IdP, PEC, {{LSConxt}pk(PEC)}sk(IdP));

//receive username from user

recv_1(User, IdP, {{uname}pk(IdP)}sk(User));

//receive attributes from user

recv_2(User, IdP, {{attrU}pk(IdP)}sk(User));

//receive attributes from SP

recv_3(SP, IdP, {{attrSP}pk(IdP)}sk(SP));

claim_idp(IdP, Secret, uname);

claim_idp(IdP, Secret, attrU);

claim_idp(IdP, Secret, attrSP);

claim_idp(IdP, Secret, LSConxt);

role PEC {

varLSConxt, rslt: Data;

freshqury, challenge: Nonce;

//PEC quiries AA for legitimate information

send_5(PEC, AA, {{qury}pk(AA)}sk(PEC));

127
Chapter 5. User-Centric Attribute-Based Access Control Model Using XACML

//PEC send challenge to user

send_7(PEC, User, {{challenge}pk(User)}sk(PEC));

//PEC receives loncal security from PEC

recv_4(IdP, PEC, {{LSConxt}pk(PEC)}sk(IdP));

//PEC receives quiry result from AA

recv_6(AA, PEC, {{rslt}pk(PEC)}sk(AA));

claim_pec(PEC, Secret, LSConxt);

claim_pec(PEC, Secret, qury);

claim_pec(PEC, Secret, rslt);

claim_pec(PEC, Secret, challenge);

role AA {

varqury: Nonce;

freshrslt: Data;

//send quiry result to AA

send_6(AA, PEC, {{rslt}pk(PEC)}sk(AA));

//receive quiries from PEC

recv_5(PEC, AA, {{qury}pk(AA)}sk(PEC));

claim_aa(AA, Secret, qury);

claim_aa(AA, Secret, rslt);

128
 Chapter 5. User-Centric Attribute-Based Access Control Model Using XACML

Figure 5.8 Validation of the Proposed Model by Scyther Toll

The result of Scyther has proved the claims of implemented protocol. During the

whole message flow, all communications is claims as secret. No potential attacks

are found. It can been seen from Figure 5.8 that all secret parameters from each

actor in the proposed model is secure.

129
 Chapter 5. User-Centric Attribute-Based Access Control Model Using XACML

5.4.2 Implementation and Tests

The proposed model is implemented in a mobile web service platform as a

prototype. The development is based on the open source software development

platform[15, 181-184]. The functionalities of four entities; IdP, SP, mobile user,

and the AA were developed separately as shown in Figure 5.5. The

communications between each entity uses SOAP messages.

Figure 5.9Proof of Concept

Client app for the mobile device is developed using Eclipse and the

implementation is based on Android JDK using version 2.3.3. The AA‟s and SP‟s

130
 Chapter 5. User-Centric Attribute-Based Access Control Model Using XACML

functionalities are developed and deployed on JAX-WS web service using Apache

Tomcat Web Server [183] in Netbeans [185]. The IdP is also built on top of the

Tomcat server. The administration functionalities are implemented using JSP

containers. The J2EE and JSP containers connect with MySQL database for storing

and retrieving of administrative information such as identity information that are

stored at the IdP‟s database. The AA also stores all the attributes in the database.

The XACML engine is developed based on Sun XACML 2.0 implementation

[182].

The architecture is feasible to integrate more than one SP into the systems and

SPs can be implemented using different software development approaches. The

implementation of access control system is based on the XACML architecture. The

following section will discuss the message standard of XACML language.

5.4.3 Sample Screenshots of the Client Application

The screenshots of Android emulator are provided in Figures5.6-5.9. These figures

depict several key steps such as user login (Figure 5.6), the SP's request for the

user‟s PII (Figure 5.7), policy evaluation process (Figure 5.8), and policy decision

(Figure 5.9) of the proposed user-centric attribute-based access control model. The

author implemented the XACML policy evaluation process as an online back-end

service. In Netbeans, the policy evaluation can be tested as a stand-alone service

under Glassfish server. The Figure 5.10 depicts the screenshot of the policy

evaluation verification page.

131
Chapter 5. User-Centric Attribute-Based Access Control Model Using XACML

Figure 5.10Login Page of the Client Application

Figure 5.11Requesting Page of the Client Application

132
Chapter 5. User-Centric Attribute-Based Access Control Model Using XACML

Figure 5.12Information Gathering at PEC

Figure 5.13Authorization Result of PEC Service

133
 Chapter 5. User-Centric Attribute-Based Access Control Model Using XACML

Figure 5.14Service Test Page

Figure 5.14 depicts the test page of the PEC service. In Figure 5.14, the

following two different methods are given: generateRequest() for generating

XACML request and pdpService() for policy evaluation. The first method is a

service for PEC to generate a dynamic XACML request for the SP on user‟s PII

data. The second method is a service of PEC to evaluate the PII access request.Let

us make the pdpService() as an example. In order to test the pdpService(), the

following four parameters need to be entered: user_id, user‟s email address, the

authentication state of the user, and the current location.

Figure 5.15 shown below is a screenshot of SOAP request and SOAP response

when a data requester is successfully granted access on user's PII. The SOAP

request part in Figure 5.15 shows the access request, which includes the four

required attributes from the data requester: user_id, email address, state of

134
 Chapter 5. User-Centric Attribute-Based Access Control Model Using XACML

authentication, and location data. The SOAP request is received by the PEP and

formatted by CH. PDP evaluates the formatted request based on the pre-defined

policies.

Figure 5.15PEC Service Test

The SOAP response part in Figure 5.15 shows the decision of the access

request. The decision is made by PDP and CH converts it to original format and

sends it back to PEP. Thus, PEP can forward the decision to the user and SP.

It shows that the XACML policy evaluation process is successfully loaded and

works fine.

135
 Chapter 5. User-Centric Attribute-Based Access Control Model Using XACML

5.4.4 Possible Extension

The implementation was currently worked under the scenario for one SP. However,

based on the study of the existing identity management system in Section 3, our

system can accommodate more than one SPs. User‟s identity can be considered as a

federated identity. Therefore, our system can be extended to accommodate a list of

SPs as long as they have done the registration process described in Section 5.1.1.

5.4.5 XACML Message Standard

This section describes polices used for the data access control, and construction of

XACML request/response messages during the transmission.

5.4.5.1 XACML Policy

At the root tag of XACML policies is a Policy or a PolicySet. A PolicySet is a

container that can hold other Policies or PolicySets, as well as references to policies

found in remote locations. A Policy represents a single access control policy,

express through a set of Rules. Each XACML policy document contains exactly

one Policy or PolicySet root XML tag. Therefore, a Policy or PolicySet may

contain multiple policies or Rules, each of which may have different access

decisions. XACML defines a collection of Combining Algorithms which reconcile

the decisions from each decision that is evaluated by policies or rules.

A policy example which defines that any user who holds an email address from

the domain of “city.ac.uk” can get access to read any resources at

“http://www.city.ac.uk” is presented in the section Appendix

136
 Chapter 5. User-Centric Attribute-Based Access Control Model Using XACML

5.4.5.2 XACML Request Context

In a XACML policy-based system, every access request has to be formatted using

XACML‟s specification. A XACML request contains the necessary information to

make authorization decisions. Basically it has the subject‟s (requestor) attributes,

the resources‟ attributes and the operation that the subject wants to perform on the

resource.

Regarding to the above policy example, a user who holds an email account

“[email protected]” requests the resources of School of Engineering and

Mathematical Sciences at http://www.city.ac.uk. This XACML request is given in

Appendix.

5.4.5.3 XACML Response Context

As a result of evaluating the policy, the response context contains only the decision

and the requested resource. Regarding the policy example and response context

example, the user has an email account “[email protected]”, hence she can obtain

the data consent on the resources at “http://www.city.ac.uk/”. The policy evaluation

decision should be Permit. The response context is listed in Appendix.

5.5 Discussion

Attributes based access control scheme, which makes access decision based on the

attributes of requestors, resources, and environment can provide greater flexibility

in today‟s environment. The model is constructed based on XACML. XACML

model use attributes as the key factor to authorize data consent. With standardised

specifications and components, it can flexibly solve privacy related issues for

137
 Chapter 5. User-Centric Attribute-Based Access Control Model Using XACML

mobile devices. In XACML based model, each access request, the complete

XACML hierarchy must be processed in order to find the matching policy sets,

policies and rules, to evaluate them ,and to compute access decision result[186].In

reality, authorization policies comprise of thousands of users and millions of

resources. A XACML based model requires service to establish all the required

components to complete a XACML request/response, which reduces the flexibility

to build a privacy preserving model for mobile devices. The efficiency of the whole

system may be decreased by such process. The computing and memory

consumption for evaluating policies are high cost tasks. XACML message is

transmitted in XML format and the parsing overhead of XML documents is also an

issue [187] that needs attention. As a result, a more light-weight and general

solution is required to satisfy today‟s security and privacy issues of mobile devices.

138
 Chapter 5. User-Centric Attribute-Based Access Control Model Using XACML

5.6 Conclusion

The proposed user-centric policy-based access control model uses the attributes

from any related stakeholders in the mobile environment. By involving the trusted

attribute authority, user has the ability to check the legality of the requested service

providers and define his/her own access privacy policies to protect the sensitive

personal information. Thus, the user has the control of his online data to decide who

can be granted access to the personal information. The access can also be restricted

to a certain part of the personal information. By using specific attributes that are

provided by the mobile devices, services providers and the related environment, the

proposed user-centric access control model establishes a secure mobile Web

service environment. In addition to this, it helps to grow the customer confidence

and also provides seamless access to their mobile financial services from anywhere,

at any time at a touch of a button.

139
6 Context-Aware Attribute-Based Encryption Schemes

6.1 Introduction

This chapter gives a design of context-aware attribute-based encryption (ABE)

scheme for smart mobile devices. Due to the recent technological advancements in

mobile devices, powerful processors and various sensors are embedded within

smartphones. This trend allows millions of people to use smartphones for their

work and day-to-day social activities. In order to support this transformation, the

traditional Cloud computing infrastructures are being modified into mobile Cloud

computing.

Chapter 5 presented a model for mobile Cloud users to enforce access control

over their PII to prevent the privacy attack from unauthorized service providers.

However, in Cloud infrastructures, platforms and applications are provided by the

third-party Cloud service providers. Hence, the Cloud service providers have

control over the data that is stored in the Cloud storage and can monitor the

communications between the end user and the Cloud with or without their prior

consent. The Cloud computing organisations should invest heavily in risk

assessment to ensure system security and compliance.

Furthermore, the user experience and usability are important factors that can

influence a mobile application. Users prefer using applications with good

performance and user interface. Traditional security mechanisms such as multiple

passwords, several memorable security questions, security patterns etc. should be

140
 Chapter 6. Context-Aware Attribute-Based Encryption Schemes

avoided in mobile Cloud environment as they are complex and infeasible. The

smartphone features should be exploited to provide robust access control with less

user operations. Therefore, the contextual attributes such as location, time, app

usage, unlock failure etc. in mobile Cloud environment should be involved in the

security framework.

This chapter presents a context-aware attribute-based encryption scheme for smart

mobile devices. It extends the functionality of the proposed model in Chapter 5 for

Cloud data storage with data confidentiality and addresses some shortcomings in

usability and user experience. The real-time contextual attributes are further

investigated via profile-behaviour profiling techniques. The framework is designed

to address the following issues:

- The confidentiality and anonymity for data storage in mobile Cloud

environment.

- Development of lightweight cryptographic technology, hence the

data owner encrypts, uploads, and decrypts the data using mobile

device with low computational cost.

- The data owner has the control of his personal data, and can define

flexible privacy access policies using context information (e.g.

location, time, date).

- The proposed framework reduces the communication overheads

compared to the conventional schemes since it is designed for

mobile networks and utilizes the existing Cloud infrastructure.

141
 Chapter 6. Context-Aware Attribute-Based Encryption Schemes

- Reduce the existing trust relationship between each stakeholder to

reduce the security risk.

The following three algorithms are proposed in the following sections: (1)

context-aware ABE scheme based on single authority, (2)context-aware ABE

scheme based on multiple authorities, and (3) low-complexity multi-authority ABE

scheme which is designed to reduce the computation and communication overheads

of mobile devices.

The available context-related attributes of mobile Cloud environment are used

in the proposed algorithms in order to strengthen the data security. The

context-related attributes can be determined by behaviour-profiling techniques[93,

188-190]. Mobile user‟s activities such as app usage, network usage, charging

times and unlock failures have been used to profile the user-behaviour. This

user-behaviour profile can be used to detect anomaly activities and provide more

fine-grained level of security.

Let us assume that there is an app installed in mobile-device which can be used

to detect anomalous activities (let us call this app as “behaviour-profiling” app).

Installing the behaviour-profiling app in the user‟s mobile device can be used to

capture the real-time attributes. Such attributes can be used to verify whether the

current user is the phisical owner of the mobile device. In the following three

algorithms, assuming that user‟s mobile device has the profiling-behaviour app

installed and the context-related attributes can be captured by the app. There is a

mapping function M, embedded within the behaviour-profiling app. M outputs “yes”

142
 Chapter 6. Context-Aware Attribute-Based Encryption Schemes

if the particular attributes satisfied the requirements. The outputs will be used by the

following three algorithms for the user to recover the encrypted data.

6.2 Context-Aware Single Authority Attribute-Based Encryption

Scheme

In this section, a context-aware attribute-based encryption scheme with single

authority is described. It is designed for mobile users to share and store their

personal data securely in the Cloud storage. A user has his personal data such as

location records, medical history stored in the Cloud. He would not allow any

unauthorized requesters to access this sensitive information, even the Cloud

service provider. This can only be achieved by storing the data in the encrypted

form in the Cloud storage. In the meantime, access control techniques should also

be deployed to restrict access to the data. Attribute-based encryption techniques

are considered as the ideal solution together with the access control models for

authentication of the users. A data owner can upload encrypted data in the Cloud,

and the encrypted data was enforced with access control polices during the

encryption process. Access policies are defined by the data owner, thus, he

determines who can access the data. A user will request data from an Attribute

Authority, which is responsible for maintaining the attributes, verifying the user

authenticity, and also issuing the decryption keys.

143
 Chapter 6. Context-Aware Attribute-Based Encryption Schemes

Figure 6.1The Framework of Single Authority ABE scheme.

Figure 6.1 shows the architecture of single authority ABE scheme. Four actors are

involved in the scheme, the data owner, the AA, the user and Cloud Storage. A

data owner uses the contextual attributes and static attributes from AA to

encrypted his data and uploads it to the Cloud storage. A user downloads the data

and requests decryption keys based on his attributes. The AA verifies the user and

issues the decryption keys to the user.

6.2.1 Preliminaries

This section gives the background information and security assumption for the

ABE scheme. Figure 6.1 shows the architecture of the single authority ABE

scheme.

144
 Chapter 6. Context-Aware Attribute-Based Encryption Schemes

6.2.1.1 Bilinear Pairings

Let 𝔾1 and 𝔾2be two cyclic groups of order q for some large prime q and g is a

generator of 𝔾1. A bilinear map e :𝔾1× 𝔾1→ 𝔾2between these two groups. The

map must satisfy the following properties:

 Bilinear: A map e :𝔾1× 𝔾1→ 𝔾2is bilinear if e(ga, gb) = e(g, g)ab for all P,

Q ∈ 𝔾1and all a, b ∈ ℤ.

 Non-degenerate: The map does not send all pairs in 𝔾1× 𝔾1to the identity

in 𝔾2. Observe that since 𝔾1, 𝔾2are groups of prime order, this implies

that e(g, g) ≠ 1.

 Computable: There is an efficient algorithm to compute e(P, Q) for any P,

Q ∈ 𝔾1.

6.2.1.2 Complexity Assumption: Decisional Bilinear Diffie-Hellman

(DBDH) Assumption

Suppose a challenger chooses a, b, c, z ∈ ℤq at random, the DBDH assumption is

that no polynomial-time adversary is to be able to distinguish the tuple (A = ga, B =

gb, C = gc,Z = e(g, g)abc) from the tuple (A = ga, B = gb, C = gc,Z = e(g, g)z) with

more than a negligible advantage.

6.2.1.3 Secret Sharing Scheme

In a secret sharing scenario, a data D is divided into n pieces. In order to reconstruct

D, a threshold value k where 1< 𝑘 ≤ 𝑛 is defined. Only if there are k pieces shares

145
 Chapter 6. Context-Aware Attribute-Based Encryption Schemes

or more comes together, then user can reconstruct the secret D. Complete

knowledge of (k – 1) pieces reveals no information about D. The scheme was firstly

invented by Adi Shamir in 1979 [161].

6.2.2 Construction

The section will begin by explain how the traditional ABE scheme can be

converted into context-aware ABE scheme. In a single authority ABE scheme,

there is only one attribute authority (AA) maintains a set of attributes and

corresponding encryption and decryption credentials. In contrast to the

conventional ABE scheme, a data owner in the proposed scheme can encrypt the

data by not only using credentials obtained from AAs but also using context-related

attributes based on the data owner‟s preferences.

The proposed algorithm is composed of four sub algorithms namely Setup,

Key Issuing, Encryption and Decryption. Functionalities of each algorithm are

briefly explained in the following:

 Setup: The setup algorithm takes a security parameter as input and

outputs a bilinear group and a set of security parameters.

 Key Issuing: AA generates decryption key for a user u that holds a

set of attributes.

 Encryption: The encryption algorithm takes a set of attributes

maintained by AAs as well as a set of context-related attributes

146
 Chapter 6. Context-Aware Attribute-Based Encryption Schemes

defined by data owner and data as inputs and outputs the ciphertext

of the data.

 Decryption: The decryption algorithm inputs the decryption

credentials received from AAs and context-related parameters

obtained from the smart mobile device and the ciphertext. Then it

outputs the original data.

The descriptions of the four sub algorithms are detailed below.

Setup S

 For a given security parameter 𝜆and 𝜎 ∈ {0, 1}𝑠 𝑝𝑜𝑙𝑦 (λ), group bilinear

parameters are generated by the AA: q, g, 𝔾1⟵S(1𝜆 ; 𝜎).

 AA randomly picks a secret y, tA,i ∈ ℤq for each attribute i(1 ≤ 𝑖 ≤ n)

and publishes corresponding public keys {Ti = 𝑔𝑡𝐴,𝑖 } ∀i and Y = e (g, g)y.

Key Issuing KI

 To issue decryption keys to user u, the AA chooses a random polynomial

pu with degree (d – 1) where pu(0) = y.

 The AA generates decryption credential for ith attribute for the user u as:

Dui = 𝑔𝑝 𝑢 (𝑖) 𝑡 𝐴 ,𝑖
where ∀i ∈ Au(Au denotes the attributes set of the user

u).

Encryption E

147
 Chapter 6. Context-Aware Attribute-Based Encryption Schemes

The data owner encrypts the message m for a set of attributes Am= AA∪AC,

where AA = {aa,1, …, aa,N} denotes the attributes maintained by AA and AC =

𝑎𝑐,1 , … , 𝑎𝑐,𝑛 denotes the context-related attributes defined by the data owner, as

follows:

 The data owner randomly chooses sA , sB∈ ℤq, and encrypts the message

as 𝐸𝑛𝐶𝑚 = m𝑌 𝑠𝐵 .

 The data owner computes E0 = h(M(aa,1)∥ M(aa,2)∥… M(aa,n))𝑌 𝑠𝐴+𝑠𝐵 , Ei =

𝑔𝑡𝑖 𝑆𝐴 ∀i ∈ AA, where h : {0, 1} ⟶ ℤq is a secure function, M is a

mapping function of context related attributes and ∥ denotes

concatenation.

 Now the data owner uploads CTm={𝐸𝑛𝐶𝑚 , E0, Ei, ∀i ∈ AA, and AC} into

the Cloud.

Decryption D

 User downloads CTm from the Cloud and checks the required attributes

to obtain the message m.

 User computes e(Ei, Di) = e(g, g)𝑝 𝑢 (𝑖)𝑆𝐴 for any attribute i∈AC∩ Au.

 Using interpolation technique, user can compute 𝑌𝑆𝐴 = e(g, g)𝑝(0)𝑆𝐴 =

e(g, g)𝑦 𝑆𝐴 .

Now the behaviour-profiling application installed in user‟s mobile device

computes the hash value of context related attributes such as location, risk-level

associated with current location and risk-level associated with user behaviour and

outputs h(M(a’a,1)∥M(a’a,2)∥…M(a’a, n)).

148
 Chapter 6. Context-Aware Attribute-Based Encryption Schemes

User can decrypt the data as follows:

ℎ(𝑀(𝑎’𝑎, 1) ∥ 𝑀(𝑎’𝑎, 2) ∥ ⋯ 𝑀(𝑎’𝑎, 𝑛))Y S A

𝐸𝑛𝐶𝑚 ∙
𝐸0

ℎ(𝑀(𝑎’𝑎, 1) ∥ 𝑀(𝑎’𝑎, 2) ∥ ⋯ 𝑀(𝑎’𝑎, 𝑛))Y S A

= 𝑚𝑌𝑆𝐵 ∙
ℎ(𝑀(𝑎𝑎, 1) ∥ 𝑀(𝑎𝑎, 2) ∥ ⋯ 𝑀(𝑎𝑎, 𝑛)) 𝑌 𝑠𝐴+𝑠𝐵

If and only if:

h(M(aa,1) ∥M(aa,2)∥…M(aa,n)) = ℎ(𝑀(𝑎’𝑎, 1) ∥ 𝑀(𝑎’𝑎, 2) ∥ ⋯ 𝑀(𝑎’𝑎, 𝑛)),

then the user can get the message m.

The novelty in our scheme compared to the conventional ABE scheme lies in

encryption and decryption sub algorithms which are detailed in the following:

denote the context related attribute set defined by the data owner as AC=

{aC,1, …,aC,n}, where aC,I denotes context related attributes. For sake of simplicity,

consider the following three context related attributes: aC,1 = “location”, aC,2 = “risk

level with his recent app usage” and aC,3 = “unlock failures in last two days”. Now

the data owner defines AC as follows:

AC = {aC,1 = “London”, aC,2< “3” and aC,3< “2”}, assume that risk level

varies between 1 to 10 where higher risk denoted by larger value. Then computes

𝐸0 = ℎ(𝐿𝑜𝑛𝑑𝑜𝑛 ∥ 𝑦𝑒𝑠 ∥ 𝑦𝑒𝑠)𝑌𝑆𝐴 +𝑆𝐵 .

This is the hash value of the required context-related information and is

considered as the access policy of the encrypted message. Using the client

149
 Chapter 6. Context-Aware Attribute-Based Encryption Schemes

application installed in the mobile device, the user will capture the real-time context

related attributes such as, location. The mapping function, M, outputs “yes” if the

current risk level is less than threshold defined by employer. This ensures that the

employee has all the credentials from AA, context-related attributes enforced by the

employer also need to be satisfied in order to decrypt the message.

The security and performance analysis will be discussed in the Section6.3.3.

6.3 Context-Aware Multi-Authority Attribute-Based Encryption

Scheme

In this section, the author describes the context-aware MA-ABE scheme. In a single

authority scenario, there is only one AA will monitor all the attributes and issues

encryption and decryption credentials for the users. It is a fully trusted party which

users have to prove his identity in order to obtain a decryption credential. For

instance, to decrypt a message m, a user proves his identity with required set of

attributes and receives the decryption credentials based on those attributes. In this

case, the AA has too much power which provides itself with the ability to decrypt

all communication messages and knows all the user‟s attributes. In the event of

corruption and/or compromise, the message‟s confidentiality cannot be achieved

and user‟s privacy can be revealed to the attackers. This is one of the drawbacks in

single authority based ABE scheme.

Meanwhile, it is more convenient to monitor and maintain different sets of

attributes by different authorities in real world. As mentioned in Section 3.6.3, NHS

manages UK national health information system and the UK drivers and vehicles‟

150
 Chapter 6. Context-Aware Attribute-Based Encryption Schemes

information is maintained by the Department of Vehicle Licensing Agency

(DVLA). Hence, in reality, it is impossible for a single authority to monitor all these

attributes and hence a viable solution is to go for multiple authorities based ABE

scheme (MA-ABE). Several MA-ABE schemes without considering context

information were proposed in [164-166].

In the context-aware MA-ABE scheme, each AA manages different sets of

attributes and issues credentials for the employees based on a set of attributes. The

AA defines a threshold value dk so that the message can be decrypted only if a user

has at least dk number of given attributes from the AA. There is no trusted central

authority exists for issuing keys and verifying identities.

6.3.1 Preliminaries

6.3.1.1 Bilinear Pairings and Complexity Assumption

The bilinear pairings and complexity assumptions are the same in section 6.2.1.

6.3.1.2 Two-phase Commit Protocol

Two-phase commit protocol (2PC protocol) is a distributed algorithm that

coordinates all the processes that participate in a distributed transaction on whether

to commit or abort the transaction [191, 192]. The protocol is a standard protocol

for making commit and abort atomic. In a normal execution of a single distributed

transaction, the protocol consists of two phases:

(1) The Commit-request Phase: In this phase, a coordinator attempts to

prepare all the required participating process for transactions. Define

151
 Chapter 6. Context-Aware Attribute-Based Encryption Schemes

necessary steps for either committing or aborting the transaction for

every participant.

(2) The Commit Phase: The coordinator decides whether to commit or

abort the transaction based on the response from the participants

(commit or abort), and notifies the results to all the participants. Then

the participants take actions regards to the result.

6.3.1.3 Anonymous Key Issuing Protocol

In our MA-ABE scheme, the central authority which holds the master secret was

removed. An anonymous key issuing protocol under the structure of Chase and

Chow‟s scheme [166] is presented.

In MA-ABE system, public parameters are available for users and AAs. A user

can request secret keys from an AA provided he has some attributes which are

managed by the AA. Then key generation algorithm will be executed by the AA

and the corresponding secret keys will be returned to the user. When a data owner

wants to encrypt a message, he makes use of the public parameters together with an

attribute set of his choice to carry out the encryption. Any user can obtain the

decryption keys from the corresponding keys to the attributes.

Define GID as the global identifier of the user. u is the GID of the user; it can

be a hash value of the GID. The user uses the anonymous key issuing protocol to

obtain decryption keys without the central authority. Before starting the key issuing

process, the kth AA has to generate several parameters.

Using the key issuing algorithm presented, the kth AA shares a secret 𝑠𝑘,𝑗 , with

𝑥
jth AA and picks 𝑥𝑘 ∈ ℤ𝑞 , and computes 𝑦𝑘 = 𝑔1 𝑘 . Thus, the kth AA and jth AA

152
 Chapter 6. Context-Aware Attribute-Based Encryption Schemes

𝑥 𝑥𝑗 𝑠𝑘,𝑗 +𝑢
can jointly compute the shared secret 𝑔1 𝑘 . The kth AA also has private

keys: α, β, γ ∈ ℤ𝑞 . Figure 6.2 depicts the flow of the anonymous key issuing

protocol. Each step in the protocol is described below:

(1) A two-phase commit protocol is used to start the key issuing and

takes u,𝜌1 from the user and 𝑠𝑘,𝑗 from the kth AA as input, and

outputs 𝑋 = 𝑠𝑘𝑗 + 𝑢 𝜌1 .

(2) The kth AA picks a random 𝜏 ∈ ℤ𝑞 , computes 𝑋1 = 𝑔𝜏 𝑥 and

𝑋2 = ℎ𝛼𝜏 .

(3) The kth AA sends 𝑋1 , 𝑋2 and proof of knowledge of the secret values

(i.e. 𝛼, 𝜏, 𝑥).

𝜌 𝜌2
(4) The user picks a random 𝜌2 ∈ ℤ𝑞 and computes 𝑌 = 𝑋1 1 𝑋2 .

Then the user sends Y and proof of knowledge of 𝜌2 to the kth AA.

(5) The kth AA computes 𝑍 = 𝑌 𝛾 𝜏

and forwards Z and proof of

knowledge of γ and τ to user.

(6) The user computes 𝐷 = 𝑍 1 𝜌2

. Therefore, user gets the key D as a

corresponding decryption key.

153
 Chapter 6. Context-Aware Attribute-Based Encryption Schemes

Figure 6.2Anonymous Key Issuing Protocol

6.3.2 Construction

Similar to single authority case, the proposed algorithm is also composed of four

sub algorithms namely as Setup, Key Issuing, Encryption and Decryption. In the

following functionalities of each algorithm are briefly explained.

 Setup: The setup algorithm takes a security parameter as input

and outputs a bilinear group and a set of security parameters. Let

154
 Chapter 6. Context-Aware Attribute-Based Encryption Schemes

us denote the total number of AAs as K and number of attributes

maintained by kth AA as 𝑛𝑘 .

 Key Issuing: Each AA generates decryption keys for a user u that

holds a set of attributes using the anonymous key issuing

protocol defined in Section 5.3.1.3.

 Encryption: The encryption algorithm takes a set of attributes

maintained by AAs as well as a set of context-related attributes

defined by data owner and data as inputs. Then it outputs the

ciphertext.

 Decryption: The decryption algorithm inputs the decryption

credentials received from each AA and context-related

parameters obtained from smart mobile device and the

ciphertext. The output will be the original data.

Let us describe the context-aware MA-ABE scheme below:

Setup S

For a given security parameter 𝜆 and 𝜎 ∈ {0, 1}𝑝𝑜𝑙𝑦 (λ) , group bilinear

parameters are generated by the AAs: q, 𝑔1 𝑔2 , 𝔾1, 𝔾2, 𝔾T⟵S(1𝜆 ; 𝜎).

Now, the AAs interact with each other and execute the following:

 kth AA randomly chooses 𝑣𝑘 ∈𝑅 ℤ𝑞 and computes 𝑌𝑘 =

𝑣𝑘
𝑒 𝑔1 , 𝑔2 , and sends 𝑌𝑘 to other AAs, where each AA computes

𝑌= 𝑌𝑘 = 𝑒 𝑔1 , 𝑔2 𝑘 𝑣𝑘 .

155
 Chapter 6. Context-Aware Attribute-Based Encryption Schemes

 Each pair of AAs shares a secret, kthAA and jth AA randomly

choose 𝑠𝑘𝑗 ∈ ℤ𝑞 , such that 𝑠𝑘𝑗 = 𝑠𝑗𝑘 .

𝑥
 kth AA randomly chooses 𝑥𝑘 ∈ ℤ𝑞 and computes 𝑦𝑘 = 𝑔1 𝑘 .

Using the share secret 𝑠𝑘𝑗 and u, kth AA and jth AA computes

𝑥 𝑠𝑘𝑗 +𝑢 𝑥𝑘 𝑠𝑗𝑘 +𝑢
𝑦𝑘 𝑗 and 𝑦𝑗 respectively.

Then each AA carries out the following steps individually:

 For ith attribute stored in the kth AA, where 𝑖 ∈ 1, … , 𝑛𝑘 , and

𝑘 ∈ 1, … , 𝐾 the kth AA randomly chooses a secret 𝑡 𝑘,𝑖 ∈ ℤ𝑞

𝑡
and computes corresponding public key as 𝑇𝑘,𝑖 = 𝑔2𝑘,𝑖 .

Key Issuing KI

 User u executes the following steps with each kth AA:

𝑅 𝑥 𝑠𝑘𝑗 +𝑢
 For 𝑗 ∈ 1, … , 𝐾 𝑘 , user gets the 𝐷𝑘𝑗 = 𝑔1 𝑘𝑗 𝑦𝑘 𝑗 for

𝑅 𝑠𝑘𝑗 +𝑢 𝑥𝑗
𝑘 > 𝑗 or 𝐷𝑘𝑗 = 𝑔1 𝑘𝑗 𝑦𝑘 if 𝑘 < 𝑗, where 𝑅𝑘𝑗 ∈ ℤ𝑞 is a

random value.

After obtaining all 𝐷𝑘𝑗 , user computes

𝑅
 𝐷𝑢 = 𝑘,𝑗 ∈ 1,…,𝐾 × 1,…,𝐾 𝑘 𝐷𝑘𝑗 = 𝑔1 𝑢 , where

𝑅𝑢 = 𝑘,𝑗 ∈ 1,…,𝐾 × 1,…,𝐾 𝑘 𝑅𝑘𝑗 .

156
 Chapter 6. Context-Aware Attribute-Based Encryption Schemes

 If user u satisfies dk number of attributes, then kth AA randomly

picks a dk-degree polynomial 𝑝𝑘,𝑢 with 𝑝𝑘,𝑢 0 = 𝑣𝑘 −

𝑗 ∈ 1,…,𝐾 𝑘 𝑅𝑘𝑗 .

𝑝 𝑖 𝑡 𝑘 ,𝑖
 kth AA computes 𝑆𝑘,𝑖 = 𝑔1 𝑘,𝑢 , 𝑖 ∈ 1, … , 𝑛𝑘 for each

eligible attribute i for the user.

Encryption E

The data owner encrypts the data m for attribute set 𝐴𝑚 = 𝐴𝐴1 ∪ 𝐴𝐴2 ∪ … 𝐴𝐴𝐾 ∪

𝐴𝐶 as follows:

 The data owner randomly picks 𝑠𝐴 , 𝑠𝐵 ∈ ℤ𝑞 and encrypts the data as

follows: 𝐸𝑛𝐶𝑚 = m𝑌 𝑠𝐵 .

The data owner then computes

𝑠 𝑠
 E0 = h(M(aa,1)∥ M(aa,2)∥… M(aa,n))𝑌 𝑠𝐴+𝑠𝐵 , 𝐸1 = 𝑔2𝐴 , 𝐶𝑘,𝑖 = 𝑇𝑘,𝑖𝐴 ,

𝑖 ∈ 𝐴𝐴𝑘 , 𝑘 ∈ 1, … , 𝐾 .

 Now the data owner uploads the encrypted data 𝐶𝑇𝑚 =

𝐸𝑛𝐶𝑚 , 𝐸0 , 𝐸1 , 𝐶𝑘,𝑖 𝑖 ∈ 𝐴𝐴 𝑎𝑛𝑑𝐴𝐶 into the Cloud.

Decryption D

The user downloads the 𝐶𝑇𝑚 from the Cloud and checks the required

attributes to decrypt the data.

157
 Chapter 6. Context-Aware Attribute-Based Encryption Schemes

For each kth AA:

 Using 𝑆𝑘,𝑖 and the corresponding 𝐶𝑘,𝑖 , user computes𝑒 𝑆𝑘,𝑖 , 𝐶𝑘,𝑖 =

𝑠𝐴 𝑝 𝑘,𝑢 𝑖
𝑒 𝑔1 , 𝑔2 .

𝑠𝐴 𝑝 𝑘,𝑢 𝑖
 User interpolates all 𝑒 𝑔1 , 𝑔2 and gets 𝑃𝑘,𝑢 =

𝑠𝐴 𝑝 𝑘,𝑢 0 𝑠𝐴 𝑣𝑘 − 𝑗 ∈ 1,…,𝑁 𝑅𝑘𝑗

𝑒 𝑔1 , 𝑔2 = 𝑒 𝑔1 , 𝑔2 𝑘 .

 User multiplies all 𝑃𝑘,𝑢 together, gets

𝑠𝐴 𝑣𝑘 −𝑅𝑢 𝑌 𝑠𝐴
𝑄 = 𝑒 𝑔1 , 𝑔2 = 𝑅 𝑢 𝑠𝐴 .
𝑒 𝑔1 ,𝑔2

Now the client application installed in user‟s mobile device computes the hash

value of context-related attributes such as current location, risk-level associated

with current location and risk-level associated with user behaviour and outputs

ℎ(𝑀(𝑎’𝑎, 1) ∥ 𝑀(𝑎’𝑎, 2) ∥ ⋯ 𝑀(𝑎’𝑎, 𝑛)).

The user can decrypt the data as follows.

ℎ(𝑀(𝑎’𝑎, 1) ∥ 𝑀(𝑎’𝑎, 2) ∥ ⋯ 𝑀(𝑎’𝑎, 𝑛))e Du , E1 Q

𝐸𝑛 𝐶𝑚 ∙
𝐸0

𝑆𝐵
ℎ(𝑀(𝑎’𝑎, 1) ∥ 𝑀(𝑎’𝑎, 2) ∥ ⋯ 𝑀(𝑎’𝑎, 𝑛))Y S A
= 𝑚𝑌 ∙
ℎ(𝑀(𝑎𝑎, 1) ∥ 𝑀(𝑎𝑎, 2) ∥ ⋯ 𝑀(𝑎𝑎, 𝑛)) 𝑌 𝑠𝐴+𝑠𝐵

If and only if:

h(M(aa,1) ∥M(aa,2)∥…M(aa,n)) = ℎ(𝑀(𝑎’𝑎, 1) ∥ 𝑀(𝑎’𝑎, 2) ∥ ⋯ 𝑀(𝑎’𝑎, 𝑛)),

then the user can get the message m.

158
 Chapter 6. Context-Aware Attribute-Based Encryption Schemes

6.3.3 Security Analysis

In this section, the security of the proposed framework is analyzed. The data

confidentiality which is the most important feature of MA-ABE scheme is firstly

analysed. Then the security analyse of key issuing protocol is presented to prove the

anonymity key issuing protocol.

6.3.3.1 Confidentiality

The proposed scheme is to be proved secure in the selective ID (SID) model. In SID,

the adversary must provide the identity he wishes to attack before receiving the

parameters of the system. It is analysed following by Sahai's method in [193].

Assume there are N number of attribute authorities. Let us denote nkto be the

number of attributes monitored by each AA. Consider the following scenario:

Setup

 The adversary sends a list of attribute sets 𝐴𝐶 = 𝐴1𝐶 … . 𝐴𝐾𝐶 , one set

for each AA. He also provides a list of corrupted AAs.

 The challenger generates parameters for the system and sends them

to the adversary.

Key Issuing

 The adversary receives all the parameters which includes the system

parameters, honest AA‟ public keys, public and secret keys of the

corrupt AAs.

 The adversary makes secret key queries as he wants the AAs follow

two rules: 1) for each GID, there must be at least one honest AA

159
 Chapter 6. Context-Aware Attribute-Based Encryption Schemes

from which the adversary has fewer than dk number of attributes,

and 2) the adversary never queries the same authority twice with the

same GID.

Challenge

 The adversary sends two messages M0 and M1.

 The challenger chooses a bit b, computes the encryption of Mb for

attribute set 𝐴𝐶 and forwards this encryption to the adversary

Guess

 The adversary outputs a guess b’ that Mb‟ has been encrypted.

 The adversary is said to succeed if he can correctly identify the

encrypted message if b = b’.

𝐸𝑛 𝐶 𝑚 𝐸𝑛 𝐶 𝑚
In our scheme, the challenge encryption 𝑏 = = =
𝑌 𝑠𝐵 𝑒 𝑔1 ,𝑔2 𝑠 𝐵 𝑘 𝑣𝑘

𝑠𝐵 𝑘 𝑣𝑘
ℎ(𝑀(𝑎’𝑎, 1) ∥ 𝑀(𝑎’𝑎, 2) ∥ ⋯ 𝑀(𝑎’𝑎, 𝑛))e Du , E1 Q
𝑒 𝑔1 , 𝑔2 ∙
𝐸0

From the algorithms in the previous section, the following three values are

known.

𝑠
 e 𝐷𝑢 , 𝐸1 = e 𝑘,𝑗 ∈ 1,…,𝐾 × 1,…,𝐾 𝑘 𝐷𝑘𝑗 , 𝑔2𝐴 ;

𝑠𝐴 𝑣𝑘 −𝑅𝑢 𝑌 𝑠𝐴
 Q = 𝑒 𝑔1 , 𝑔2 = 𝑅 𝑢 𝑠𝐴 ;
𝑒 𝑔1 ,𝑔2

160
 Chapter 6. Context-Aware Attribute-Based Encryption Schemes

 ℎ(𝑀(𝑎’𝑎, 1) ∥ 𝑀(𝑎’𝑎, 2) ∥ ⋯ 𝑀(𝑎’𝑎, 𝑛));

These three values are the key parameters to decrypt the encrypted message. If

the adversary can successfully compute these values, the algorithm can be

considered as an insecure algorithm. The possibility for the adversary to get these

values is investigated.

Given that:

𝑠
e 𝐷𝑢 , 𝐸1 = e 𝐷𝑘𝑗 , 𝑔2𝐴
𝑘,𝑗 ∈ 1,…,𝐾 × 1,…,𝐾 𝑘

𝑅 𝑥 𝑠𝑘𝑗 +𝑢 𝑅 𝑠𝑘𝑗 +𝑢 𝑥𝑗
where𝐷𝑘𝑗 = 𝑔1 𝑘𝑗 𝑦𝑘 𝑗 for 𝑘 > 𝑗 or 𝐷𝑘𝑗 = 𝑔1 𝑘𝑗 𝑦𝑘 if 𝑘 < 𝑗 ,

𝑠
where 𝑅𝑘𝑗 ∈ ℤ𝑞 and 𝐸1 = 𝑔2𝐴 is a public key.

For kth AA, the adversary queries each AA for the secret. Assume jth (1 < 𝑗 <

𝑅 𝑥 𝑠𝑘𝑗 +𝑢
𝑘) AA receives the request. jth AA issues 𝐷𝑘𝑗 = 𝑔1 𝑘𝑗 𝑦𝑘 𝑗 .

𝑅 𝑥 𝑠𝑘𝑗 +𝑢
In 𝐷𝑘𝑗 = 𝑔1 𝑘𝑗 𝑦𝑘 𝑗 , the 𝑠𝑘𝑗 is the secret shared between jth and kth AA.

Since there is only one honest AA, it will exchange the secret with all other corrupt

𝑥 𝑠𝑘𝑗 +𝑢 𝑥𝑘 𝑠𝑗𝑘 +𝑢
authorities by computing 𝑦𝑘 𝑗 and 𝑦𝑗 during the Setup stage.

The 𝑅𝑘𝑗 ∈ ℤ𝑞 is issued by kth AA during the Key Issuing stage. The

adversary can get the value of 𝐷𝑘𝑗 . This is a potential risk of the MA-ABE scheme

and it will be studied in Section 6.3.3.2.

161
 Chapter 6. Context-Aware Attribute-Based Encryption Schemes

𝑠𝐴 𝑣𝑘 −𝑅𝑢 𝑌 𝑠𝐴
For Q = 𝑒 𝑔1 , 𝑔2 = 𝑅 𝑠 , let us recall that the kth AA issues a
𝑒 𝑔1 𝑢 ,𝑔2𝐴

𝑝 𝑖 𝑡 𝑘,𝑖
secret key 𝑆𝑘,𝑖 = 𝑔1 𝑘,𝑢 for the eligible attributes of user u. 𝑝𝑘,𝑢 𝑖 is a

polynomial that can be computed as a linear combination of d known points. A (d –

1) degree polynomial can be solved by d number of points on the polynomial. For

the kth AA, if the number of user‟s eligible attributes is greater than the threshold dk.

Then the user can get 𝑝𝑘,𝑢 0 .

In order to obtain Q, for any dk attributes 𝑖 ∈ 𝐴𝑘𝐶 ∩ 𝐴𝑘𝑢 at kth AA, three steps are

executed:

𝑠𝐴 𝑝 𝑘,𝑢 𝑖
Computes𝑒 𝑆𝑘,𝑖 , 𝐶𝑘,𝑖 = 𝑒 𝑔1 , 𝑔2 .

𝑠𝐴 𝑝 𝑘,𝑢 𝑖
The kth AA interpolates all the values 𝑒 𝑔1 , 𝑔2 together to get

𝑠𝐴 𝑝 𝑘,𝑢 0 𝑠𝐴 𝑣𝑘 − 𝑗 ∈ 1,…,𝐾 𝑅𝑘𝑗

𝑃𝑘,𝑢 = 𝑒 𝑔1 , 𝑔2 = 𝑒 𝑔1 , 𝑔2 𝑘 . Q can be obtained by

multiplying all 𝑃𝑘,𝑢 .

For the adversary, all the corrupt AAs issue decryption keys for him. However,

for the honest AA, the adversary with insufficient attributes cannot compute the

value of 𝑃𝑘,𝑢 because he is unable to get the value 𝑝𝑘,𝑢 0 . Thus Q is not available

for the adversary.

For the last the value of ℎ(𝑀(𝑎’𝑎, 1) ∥ 𝑀(𝑎’𝑎, 2) ∥ ⋯ 𝑀(𝑎’𝑎, 𝑛)). With the

behaviour-profiling app, user‟s context-related attributes can be captured at runtime

and the necessary operations executed during decryption. Therefore, the mobile

operating system should have a secure mechanism to protect the app and the data

from being modified.

162
 Chapter 6. Context-Aware Attribute-Based Encryption Schemes

The recently introduced robust security software such as KNOX [194] and

BES [195] virtually divided the mobile device into two isolated containers:

personal user container and work container. The apps in one container are separated

from the apps in the other container. The apps in work container are also virtually

located in the user mobile device management server. Thus, the behaviour-profiling

app can be verified by the device management server to prevent from being

modified. Thus, modifying the behaviour-profiling app by malicious user in order

to feed false results for the context-related attributes can be easily detected.

From the analysis above, it shows that if the adversary request secret keys from

an authority which he has insufficient attributes (the honest authority), the

adversary cannot output a correct guess of b’ such that b’ = b. Our scheme is secure

based on the selective ID model.

Hence, our proposed context-aware single authority ABE scheme and

context-aware MA-ABE scheme is secure in the selective ID model.

6.3.3.2 Key Issuing Analysis

In the anonymous key issuing protocol, the kth AA will set up several parameters

before staring the key issuing. Then it computes 𝑋 = 𝑠𝑘𝑗 + 𝑢 𝜌1 , 𝑋1 = 𝑔𝜏 𝑥 ,

𝑥
and 𝑋2 = ℎ𝛼𝜏 , where 𝑔 = 𝑦𝑗 𝑘 , ℎ = 𝑔1 . g and h are unknown to the user..

𝑥𝑘 𝜏 𝑠𝑘𝑗 + 𝑢 𝜌 1
Now adversary can get that 𝑋1 = 𝑔𝜏 𝑥
= 𝑦𝑗 and 𝑋2 = ℎ𝛼𝜏 =

𝑔1𝛼𝜏 where 𝛼 = 𝑅𝑘𝑗 𝑤ℎ𝑒𝑛 𝑘 > 𝑗 𝑎𝑛𝑑 𝛼 = −𝑅𝑘𝑗 𝑤ℎ𝑒𝑛 𝑘 < 𝑗.

163
 Chapter 6. Context-Aware Attribute-Based Encryption Schemes

𝜌 𝜌2 𝜌2 𝑥𝑘 𝜏 𝑠𝑘𝑗 + 𝑢 𝛼𝜏 𝜌 2
The challenger then computes 𝑌 = 𝑋1 1 𝑋2 = 𝑦𝑗 𝑔1 with a

𝛾𝜌 2 𝑥 𝑘 𝑠𝑘𝑗 + 𝑢 𝛾𝛼 𝜌 2
random𝜌2 ∈ ℤ𝑞 . The kth authority computes the 𝑍 = 𝑌 𝛾 𝜏 = 𝑦𝑗 𝑔1 .

Finally the user gets the key using

𝛾𝑥 𝑘 𝑠𝑘𝑗 + 𝑢 𝛾𝛼 𝛾𝑥 𝑘 𝑥 𝑗 𝑠𝑘𝑗 + 𝑢 𝛾𝑅𝑘𝑗

𝐷𝑘𝑗 = 𝑍 1 𝜌2
= 𝑦𝑗 𝑔1 = 𝑔1 𝑔1 ,where γ = 1 when k >

j and γ = -1 when k < j.

Note that during the kth AA issuing a key, the other AAs will also participate to

compute the pseudorandom values: the shared secret 𝑠𝑘𝑗 and the public key

𝑥
𝑦𝑗 = 𝑔1 𝑗 .

𝑥𝑘
However, values of discrete logarithm between 𝑔 = 𝑦𝑗 = 𝑔 𝑥 𝑗 𝑥 𝑘 𝑎𝑛𝑑 ℎ = 𝑔1

should be unknown. This also requires that the collusion group authorities cannot

learn the other honest authority‟s private key xj.

If there is only one honest AA, the adversary can extract the master secret such

as 𝑠𝑘𝑗 ,during the key issuing process because of the incorporation of some

pseudorandom values from other corrupt authorities. Then the protocol is not

secure.

𝑥
For kth corrupted authority, the adversary obtains 𝑦𝑗 = 𝑔1 𝑗 , from the honest jth

AA (1 < 𝑗 < 𝑘). The value is a pseudorandom value based on the shared secret

𝑠𝑘𝑗 and u. It also incorporates some pseudorandom values to get 𝐷𝑘𝑗 =

𝑅 𝑥 𝑠𝑘𝑗 +𝑢
𝑔1 𝑘𝑗 𝑦𝑘 𝑗 . If there are only one honest authority, recall that each pair of

authorities shares the 𝑠𝑘𝑗 , with the knowledge of 𝑦𝑗 , 𝑠𝑘𝑗 , 𝑢, 𝑎𝑛𝑑 𝑅𝑘𝑗 the adversary

can easily compute 𝐷𝑘𝑗 of the honest authority as a result. Thus, the secret value

164
 Chapter 6. Context-Aware Attribute-Based Encryption Schemes

e 𝐷𝑢 , 𝐸1 is revealed to the adversary. All the other corrupt AAs are able to

compute the pseudorandom values to compromise the system.

In this scenario, therefore, at least one more AAs are required to adjust its

pseudorandom to compensate. This authority must be an honest AA that the

adversary cannot request sufficient values.

Therefore, it can be concluded that our proposed context-aware MA-ABE

scheme maintains the collusion resistance against up to ( 𝑁 − 2 ) Attribute

Authorities.

6.3.4 Performance Analysis

In this section, computation and communication costs associated for both the single

and multi-authority algorithms proposed in Section 6.2 and 6.3 are analysed. The

efficiency of the proposed algorithms is demonstrated by comparing them against

the related conventional ABE schemes. The single authority ABE scheme is

discussed first followed by the discussion of MA-ABE scheme.

6.3.5 Computational Complexity Analysis

In single authority ABE scheme, the user is only involved in the computation

process during the Decryption stage and the data owner involves computation

process in the Encryption stage. The computational cost involved in the Setup and

Key Issuing stages can be ignored since those can be done in the idle time. The

computational cost of hash function is negligible compared to pairing and

165
 Chapter 6. Context-Aware Attribute-Based Encryption Schemes

exponentiation. Let us denote the computational time for one multiplication, one

exponentiation, and one bilinear pairing as 𝐶𝑚 , 𝐶𝑒𝑥 , 𝑎𝑛𝑑 𝐶𝑝 , respectively. Denote

the total number of attributes used for encryption as n and the total number of

context-related attributes used by the data owner as d. Table 6.1 shows the total

time required for encryption and for decryptions for the proposed scheme and

conventional scheme.

Conventional ABE Scheme Proposed Scheme

Encryption (n + 1)𝐶𝑒𝑥 + 𝐶𝑚 (n + 2)𝐶𝑒𝑥 + 2𝐶𝑚

Decryption n𝐶𝑝 + n𝐶𝑚 n𝐶𝑝 + (n + 2)𝐶𝑚

Table 6.1 Comparison of Computational Cost for the Single Authority ABE

Scheme and the proposed Context-Aware Single Authority ABE Scheme

From Table 6.1, it shows that in the single authority scenario, our scheme takes just

one more exponentiation operation and one more multiplication than the

conventional ABE scheme. The decryption time of the proposed scheme is

increased only by one multiplication time and independent of the number of

context-related attributes. The proposed scheme adds an additional layer of

security for negligible computational expense.

Table 6.2 compares the computational complexity of the proposed MA-ABE

scheme with Chase and Chow‟s scheme [166]. In the multi-authority scenario, the

users are involved in the Decryption stage while the data owner is involved in the

Encryption stage. For simplicity, denote the total number of AAs in the system as K

166
 Chapter 6. Context-Aware Attribute-Based Encryption Schemes

and the data owner uses n number of attributes from each AA for encryption. Let us

use the same benchmark time values given for jPBC library for comparison[196].

Table 6.3 shows the time values (in ms) for 𝐶𝑚 , 𝐶𝑒𝑥 , and 𝐶𝑝 based on two different

test beds:

(1) Intel (R) Core (TM) 2 Quad CPU Q6600 with 2.50GHz and 3GB

memory running on the Ubuntu 10.04

(2) HTC Desire HD A9191running on Android 2.2.

The time values given in Table 6.3 are for a symmetric elliptic curve called

Type A curve which has a base field size of 512-bit and the embedding degree is 2.

The type A curve has a 160-bit group order. Let us assume that the data owner uses

an environment similar to the Test bed 1 for encryption while the user uses a mobile

device similar to Test bed 2 for decryption.

Conventional ABE Scheme Proposed Scheme

Encryption (nK + 2)𝐶𝑒𝑥 + 𝐶𝑚 (nK + 3)𝐶𝑒𝑥 + 2𝐶𝑚

Decryption (nK+1)𝐶𝑝 + (nK + 1)𝐶𝑚 (nK + 1 )𝐶𝑝 + (nK + 3)𝐶𝑚

Table 6.2 Comparison of Computational Complexity of Proposed MA-ABE

Scheme with Chase and Chow‟s Scheme

167
 Chapter 6. Context-Aware Attribute-Based Encryption Schemes

Test Bed 1 (ms) Test Bed 2 (ms)

𝐶𝑝 14.6 491.2

𝐶𝑒𝑥 2.8 34.1

𝐶𝑚 1.8 20

Table 6.3 Time Complexity Measures for Two Different Test Beds

Figure 6.3Comparison of Computational Costs for Encryption Between

Conventional MA-ABE Scheme and the proposed Context-Aware MA-ABE

Scheme

168
 Chapter 6. Context-Aware Attribute-Based Encryption Schemes

Figure 6.3 shows the computational complexity in terms of total time required for

the data owner and the user to encrypt and decrypt the data, respectively for

conventional MA-ABE scheme and the proposed scheme when K = 1. For

encryption, our scheme consumes nearly 5msmore than the conventional ABE.

However, the proposed scheme incorporates the context-related attributes during

the encryption which provides runtime security to the data owner. It is worth noting

that from Figure 6.3 that the time difference between our scheme and the

conventional ABE for encryption is independent of the number of attributes (i.e.

time complexity orders for both schemes are same for encryption). However, our

scheme is capable of including context–related attributes on top of the regular

attributes. For decryption, it is obvious from Figure 6.3 that our scheme performs

equally well as the conventional ABE scheme.

Figure 6.4Comparison of Computational Costs for Decryption Between

Conventional MA-ABE Scheme and the Proposed MA-ABE Scheme

169
 Chapter 6. Context-Aware Attribute-Based Encryption Schemes

Figure 6.4 compares the decryption costs for both the proposed scheme and the

conventional MA-ABE scheme in terms of time complexity for different number of

AAs (i.e. K = 2, 4, 6, 8, 10).Encryption and decryption time increases with the total

number of AAs. For encryption, similar to K = 1 case, the time complexity orders

of both the schemes are same (i.e. our scheme consume nearly 5ms more than the

conventional MA-ABE scheme irrespective of number of attributes and number of

AAs). Moreover, for decryption, our scheme performs equally well as the

conventional MA-ABE scheme regardless of number of AAs involved in the

encryption. As seen from both Figure 6.3 and Figure 6.4, time complexity for

decryption is nearly 100 times more than the encryption due to the limited process

power at the mobile device. The time complexity can be reduced if less number of

attributes are used for encryption. The proposed scheme enables the data owner to

reduce the time complexity at the user end by reducing the number of attributes

from the AA. However, our scheme adds an extra layer of security by adding the

context-related attributes with negligible increment in complexity.

The proposed scheme enables the data owner to reduce the time complexity at

the user end by adding more context-related attributes with negligible increment in

complexity. For instance, the data owner can include five attributes from AAs and

another five context-related attributes which almost reduces the complexity by half

compared with the conventional MA-ABE scheme. However, the proposed scheme

has additional security in terms of context-related attributes which is not possible in

conventional ABE scheme. In a nutshell, the proposed scheme does not degrade the

performance of conventional ABE scheme while including context-related

170
 Chapter 6. Context-Aware Attribute-Based Encryption Schemes

attributes to enhance the security of the employer‟s data while reducing the

complexity at user‟s end.

6.3.6 Communication Complexity Analysis

The communication costs for the proposed schemes and the conventional schemes

are relying on the Key Issuing stage and when uploading and downloading the data.

Since, Key Issuing stage purely dependent on communication between authorities

and the data owner. Communication costs for both the schemes in Key Issuing

Stage are equal. During uploading and downloading stage, only the extra

components added to the proposed scheme compared to the conventional ABE

scheme are E0 = h(M(aa,1)∥M(aa,2)∥…M(aa,n))𝑌 𝑠𝐴+𝑠𝐵 , which requires 160-bits and

𝐴𝐶 (3-bits are enough to represent one context-related attribute out of eight

context-related attributes).Overall, increment in the communications cost in the

proposed algorithm is negligible.

From the performance analysis in the previous section, it shows that the

workload of the computation and communications are heavy for the mobile users.

For the communication, each user has to communicate with all the authorities for

requesting keys and execute required computations in order to obtain the

decryption keys and data. These interactions and computational works are a burden

for the mobile user in terms of communication and computation complexity. The

reliability of mobile data network is also a challenge for completing all required

communications. The next section will present a solution to address these issues.

171
 Chapter 6. Context-Aware Attribute-Based Encryption Schemes

6.4 Low-Complexity Multi-Authority Attribute-Based Encryption

Scheme

This section presents the low-complexity MA-ABE scheme which substantially

reduces the communication and computation workload to the user compared to

both the conventional MA-ABE scheme and the proposed context-aware MA-ABE

scheme presented in Section 6.3.

Since mobile Cloud computing serves mobile devices anywhere, anytime

through the channel of Ethernet or Internet regardless of heterogeneous

environments and platform [197]. The author introduced a Cloud server based

semi-trusted-authority (STA) between the user and the AAs. The user equipped

with smart devices only provides pseudonym of his identity to the STA. Then the

STA interacts with all the AAs on-behalf of the user and obtains decryption-keys.

Later, STA combines all the keys to obtain a key and pass it to the user. The user

has the ability to do necessary computations on the received keys and obtains the

final decryption keys to recover the data. Since all the distributed keys provided by

the AAs are masked, STA cannot decrypt the data. Moreover, STA cannot pool all

the keys and obtain attributes of the user. Hence, our algorithm preserves the

security and privacy of Chase and Chow‟s MA-ABE scheme while outsourcing the

computational and communication overheads to the STA.

Assume that the STA will execute the protocol correctly that he will behave in

a semi-honest manner, e.g. he is honest but curious so privacy is a real issue.

172
6.4.1 Constructions

There are five different parties involved in the proposed framework: the data owner

(encryptor), the user (decryptor), the AA, STA and the Cloud storage server. Figure

6.5 depicts the main framework and work flow of the proposed system.

Figure 6.5Illustration of the proposed Framework of Low-Complexity

Context-Aware Multi-Authority ABE Scheme for Mobile Cloud

Environment

173
 Chapter 6. Context-Aware Attribute-Based Encryption Schemes

The algorithm is composed of four sub algorithms which are named as: Setup,

Key Issuing, Encryption and Decryption. In the following functionalities of each

algorithm are briefly explained.

 Setup: The setup algorithm takes a security parameter as input and

outputs a bilinear group and a set of security parameters.

 Key Issuing: AAs generate decryption keys for a user u that holds a

set of attributes.

 Encryption: The encryption algorithm takes a set of attributes

maintained by AAs as well as a set of context-related attributes

defined by data owner as inputs and it outputs the ciphertext.

 Decryption: The decryption algorithm takes as input the decryption

credentials received from AAs and context-related parameters

obtained from smart mobile device and the ciphertext. The output

will be the original data.

Define each algorithm as follows:

Setup S

For a given security parameter 𝜆 and 𝜎 ∈ {0, 1}𝑝𝑜𝑙𝑦 (λ) , group bilinear

parameters are generated by the AA: q, 𝑔1 , 𝑔2 , 𝔾1,𝔾2,𝔾T⟵S(1𝜆 ; 𝜎).

Now, the AAs interact with each other and execute the following:

174
 Chapter 6. Context-Aware Attribute-Based Encryption Schemes

 kth AA randomly chooses 𝑣𝑘 ∈𝑅 ℤ𝑞 and computes 𝑌𝑘 =

𝑣𝑘
𝑒 𝑔1 , 𝑔2 , and sends 𝑌𝑘 to other AAs, where each AA computes

𝑌= 𝑌𝑘 = 𝑒 𝑔1 , 𝑔2 𝑘 𝑣𝑘 .

 Each pair of AAs shares a secret, kth AA and jth AA randomly

choose 𝑠𝑘 ∈ ℤ𝑞 , such that 𝑠𝑘𝑗 = 𝑠𝑗𝑘 .

𝑥
 kth AA randomly chooses 𝑥𝑘 ∈ ℤ𝑞 and computes 𝑦𝑘 = 𝑔1 𝑘 .

Using the shared secret 𝑠𝑘𝑗 and u, kth AA and j computes

𝑥 𝑠𝑘𝑗 +𝑢 𝑥𝑘 𝑠𝑗𝑘 +𝑢
𝑦𝑘 𝑗 and 𝑦𝑗 respectively.

Then each AA carries out the following steps individually:

 For ith attribute stored in the kth AA, where 𝑖 ∈ 1, … , 𝑛𝑘 , and

𝑘 ∈ 1, … , 𝐾 the kth AA randomly chooses a secret 𝑡 𝑘,𝑖 ∈ ℤ𝑞

𝑡
and computes corresponding public key as 𝑇𝑘,𝑖 = 𝑔2𝑘,𝑖 .

Key Issuing KI

The STA executes the following steps with kth AA on behalf of user u, hence

the following communication and computational overheads have been offloaded to

STA. In order to mask the decryption keys to the STA, a pre-shared secret between

user and kth AA is used when issuing keys, the key will be combined into the

pseudorandom and only the user can derive the decryption key.

𝑅 𝑥 𝑠𝑘𝑗 +𝑢
 For 𝑗 ∈ 1, … , 𝐾 𝑘 , STA gets the 𝐷𝑘𝑗 = 𝑔1 𝑘𝑗 𝑦𝑘 𝑗 for

𝑅 𝑠𝑘𝑗 +𝑢 𝑥𝑗
𝑘 > 𝑗 or 𝐷𝑘𝑗 = 𝑔1 𝑘𝑗 𝑦𝑘 if 𝑘 < 𝑗, where 𝑅𝑘𝑗 ∈ ℤ𝑞 is a

random value.

175
 Chapter 6. Context-Aware Attribute-Based Encryption Schemes

After obtaining all 𝐷𝑘𝑗 , STA computes

𝑅
 𝐷𝑢 = 𝑘,𝑗 ∈ 1,…,𝐾 × 1,…,𝐾 𝑘 𝐷𝑘𝑗 = 𝑔1 𝑢 , where

𝑅𝑢 = 𝑘,𝑗 ∈ 1,…,𝐾 × 1,…,𝐾 𝑘 𝑅𝑘𝑗 .

 If user u satisfies dk number of attributes, then kth AA randomly

picks a (𝑑𝑘 − 1 ) degree polynomial 𝑝𝑘,𝑢 ∙

 If the user doesn‟t satisfy dk number of attributes, the kth AA

randomly picks a polynomial 𝑝𝑘,𝑢 ∙ with degree nk+1, where

nkdenotes the number of attributes managed by kth AA.

 Now, kth AA use the pre-shared secret, 𝑟𝑘 between user u and kthAA,

define 𝑝𝑘,𝑢 0 = 𝑣𝑘 + 𝑟𝑘 − 𝑗 ∈ 1,…,𝐾 𝑘 𝑅𝑘𝑗 .

𝑝 𝑖 𝑡 𝑘 ,𝑖
 The kth AA computes 𝑆𝑘,𝑖 = 𝑔1 𝑘,𝑢 ,𝑖 ∈ 1, … , 𝑛𝑘 for each

eligible attribute i for the user.

Encryption E

 The data owner encrypts the message m for attribute set Ac=

{𝐴1𝐶 , …, 𝐴𝐾𝐶 } The data owner picks 𝑠 ∈ ℤ𝑞 and outputs 𝐸0 =

𝑚𝑌𝑠, 𝐸1=𝑔2𝑠, 𝐶𝑘,𝑖= 𝑇𝑘,𝑖𝑠𝑖∈𝐴𝐶𝑠, ∀𝑘∈1, …, 𝐾.

 Now the data owner uploads the encrypted data and related

parameters to the cloud storage.

176
 Chapter 6. Context-Aware Attribute-Based Encryption Schemes

Decryption D

The decryption stage has two steps. Firstly STA obtains the masked keys from

all the authorities and combine them together. Then the combined key will be

forwarded to the user to get the unmasked the combined key.

 Decryption by STA

o For each authority k:

(1) For any attribute 𝑖 ∈ 𝔸𝑘𝐶 ∩ 𝔸𝑘𝑢 , STA computes

𝑠𝑝 𝑘,𝑢 i
𝑒 𝑆𝑘,𝑖 , 𝐶𝑘,𝑖 = 𝑒 𝑔1 , 𝑔2 .

𝑠𝑝 𝑘,𝑢 i
(2) STA interpolates all 𝑒 𝑔1 , 𝑔2 together and gets

𝑠𝑝 𝑘,𝑢 0
𝑃𝑘 = 𝑒 𝑔1 , 𝑔2 =

𝑠 𝑣𝑘 + 𝑟 𝑘 − 𝑗 ∈ 1,…,𝐾 𝑅𝑘𝑗
𝑒 𝑔1 , 𝑔2 𝑘 .

o STA multiplies all 𝑃𝑘 together, and gets 𝑄=

𝑠 𝑣𝑘 + 𝑟 𝑘 −𝑠𝑅𝑢 𝑌 𝑠+𝑠 𝑟 𝑘
𝑒 𝑔1 , 𝑔2 = 𝑅 .
𝑒 𝑔1 𝑢 , 𝑔2𝑠

𝑅
o Then STA computes 𝑇 = 𝑒 𝐷𝑢 , 𝐸1 ∙ 𝑄 = 𝑒 𝑔1 𝑢 , 𝑔2𝑠 ∙ 𝑄 =

𝑌 𝑠+𝑠 𝑟𝑘
and forwards 𝑇 to the user.

 Decryption by User

o User holds the pre-shared secret 𝑟𝑘 , computes

𝑟
𝑒 𝑔1 𝑘 , 𝑔2𝑠 = 𝑒 𝑔1 , 𝑔2 𝑠 𝑟𝑘
= 𝑌𝑠 𝑟𝑘
.

o In order to recover the message, the user computes 𝑌 𝑠 =

𝐸0
𝑇 𝑌𝑠 𝑟𝑘
, recover m as 𝑚 = .
𝑌𝑠

177
 Chapter 6. Context-Aware Attribute-Based Encryption Schemes

6.4.2 Security Analysis

This section analyses the security of the proposed low-complexity MA-ABE

scheme. As the proposed scheme is an extension of Chase and Chow‟s MA-ABE

scheme [166], the security analysis showed that the proposed scheme does not

degrade the security and privacy of the encrypted message and mobile user

compared to original scheme. It satisfies data confidentiality of encrypted data

against unauthorized users and the curious Cloud service providers under the

selective ID model. It maintains the collusion resistance against up to (N – 2) AAs.

The user‟s privacy is also protected by the anonymous key issuing protocol.

The stages that differ between the two schemes are focused on, the Key Issuing

stage and the Decryption stage.

During the Decryption stage, the STA performs the steps in place of the

authority AA. In more detail, the STA only computes 𝑇 = 𝑒 𝐷𝑢 , 𝐸1 ∙ 𝑄 =

𝑅
𝑒 𝑔1 𝑢 , 𝑔2𝑠 ∙ 𝑄 = 𝑌 𝑠+𝑠 𝑟𝑘
in contrast to𝑌 𝑠 that is computed by the authority in

the Chase and Chow‟s scheme. As the required decryption key to decrypt the

message m is 𝑌 𝑠 , the STA cannot decrypt to obtain the message m, therefore the

confidentiality of the message is ensured. More precisely, since the shared secret

𝑟𝑘 is only known between the kth AA and the mobile user, and thus the summation

𝑘 𝑟𝑘 can only be obtained by a mobile user; therefore the STA cannot obtain 𝑌 𝑠

for its know expression of 𝑇 = 𝑌 𝑠+𝑠 𝑟𝑘

.

178
 Chapter 6. Context-Aware Attribute-Based Encryption Schemes

During the Key Issuing stage, the STA performs most of the steps in place of

𝑝 𝑖 𝑡 𝑘,𝑖
the user in Chase and Chow‟s scheme. The kth AA computes 𝑆𝑘,𝑖 = 𝑔1 𝑘,𝑢 ,𝑖 ∈

1, … , 𝑛𝑘 and sends them to STA. If the user satisfies the minimum 𝑑𝑘 number

of attributes, then the degree of polynomial chosen by AA is equal to 𝑑𝑘 . Hence,

𝑑𝑘 number of 𝑆𝑘,𝑖 can be used to get the secret

𝑝 0 = 𝑣𝑘 + 𝑟𝑘 − 𝑗 ∈ 1,…,𝐾 𝑘 𝑅𝑘𝑗 during the interpolation. If the user does not

satisfy the minimum 𝑑𝑘 number of attributes then the user cannot obtain enough

number of key 𝑆𝑘,𝑖 to recover the secret 𝑝 0 . This is the crucial point that for the

polynomial with degree (𝑑𝑘 − 1). A number of 𝑑𝑘 points are required in order to

get 𝑝 0 . If the user with insufficient attributes, then the kth AA chooses a

polynomial with degree (𝑛𝑘 + 1). Therefore, the STA is not able to pool all 𝑆𝑘,𝑖

from all AAs in order to find attributes of mobile user, and cannot distinguish which

set of attributes belongs to the mobile user. This also preserves the privacy of the

user.

6.4.3 Performance Analysis

In Chase and Chow‟s scheme, the user takes part in the computation during the

Key Issuing and Decryption stages. Most of the required computation and

communications take place in these two stages. Our proposed scheme successfully

offloads these time-cost operations to the semi-trusted Cloud based server. Since

the Cloud servers have powerful computing capacity and reliable network, utilizing

these powerful resources is an ideal way.

179
 Chapter 6. Context-Aware Attribute-Based Encryption Schemes

6.4.3.1 Reduction in Communication Overhead

The communication overhead is always an important factor for mobile Cloud

environment. It is important for the service performance and user experience. In

Chase and Chow‟s scheme, user needs to start (𝑁 − 1)independent invocations for

each AA during the key issuing stage. In order to compute the decryption key 𝐷𝑢 , a

user has to request the key 𝐷𝑘𝑗 from all the authorities. This process requires a

large number of communications. The more authorities requested, the more

communications take place. It is a big challenge for mobile data networks to

complete the process. Our proposed scheme offloads these communications to the

Cloud based server, so that the communication overheads can be avoided at the

user‟s end. Figure 6.6 shows the comparisons of number of communications

between the conventional MA-ABE scheme and our proposed scheme. From

Figure 6.6, the number of communications in Chase and Chow's scheme goes up

much more than our proposed scheme. In our scheme, the user only needs to do

communications with the STA, which improves the communication performance.

180
 Chapter 6. Context-Aware Attribute-Based Encryption Schemes

Figure 6.6Comparisons of Communications between Chase and Chow’s

MA-ABE Scheme and proposed Low Complexity MA-ABE Scheme

6.4.3.2 Reduction in Computational Overhead

In both the conventional MA-ABE scheme and proposed low complexity MA-ABE

scheme, the user (data requester) has to execute several computations which are

required in the Key Issuing and Decryption stages in order to decrypt the data. The

computational time (in ms) for one multiplication, one exponentiation and one

pairing as 𝐶𝑚 , 𝐶𝑒𝑥 , 𝐶𝑝 , respectively. Let us also denote the total number of

attributes of each authority n and total number of attributes used by the data owner

as d, and N denotes the total number of authorities. The following table shows a

comparison of the number of required computations between the conventional

MA-ABE scheme and the proposed scheme. From the results, it shows that for both

181
 Chapter 6. Context-Aware Attribute-Based Encryption Schemes

Key Issuing and Decryption stages, the high-cost computational tasks are

substantially offloaded to the Cloud-based server.

Conventional MA-ABE Proposed Low Complexity

Scheme MA-ABE Scheme

Key Issuing N(N-1)𝐶𝑚 _______

Decryption (Nn+1)𝐶𝑝 + (Nn + 1)𝐶𝑚 N𝐶𝑝 + (N+1)𝐶𝑚

Total (Nn+1)𝐶𝑝 +(N2 + Nn +1)𝐶𝑚 N𝐶𝑝 + (N+1)𝐶𝑚

Table6.4Comparison of Number of Required Computations between

the Conventional MA-ABE Scheme and Proposed scheme

182
 Chapter 6. Context-Aware Attribute-Based Encryption Schemes

6.5 Conclusion

The concept of ABE scheme has been demonstrated to perform a step closer in

securely sharing the data. The traditional ABE is based on a single authority which

manages all the attributes, communications and also issues the decryption keys.

From the security aspect, if the authority was attacked, the whole system will be

compromised. From the performance aspect, the authority is designed with too

much responsibility. It can be the bottleneck of the whole system when a large

number of users request keys at the same time. Furthermore, in the mobile Cloud

environment, the available contextual attributes should be used to strengthen the

security and data access control due to mobility.

In this chapter first a context-aware single authority ABE scheme is proposed

which extended the traditional ABE scheme by incorporating contextual attributes

in the mobile Cloud environments. Thus, the decryption keys are not only issued

based on the attributes that are maintained by AAs, but also rely on policies defined

by the data requestor. Following this, a context-aware MA-ABE scheme which is

more practical is presented. In reality, attributes are managed by different AAs. A

user requests the decryption keys from different authorities. The proposed scheme

also removes the dependency on the central authority which issues final decryption

keys and monitors all the communications to reduce the risk level of the system.In

this way, users can combine all the keys received from all the AAs and the real-time

contextual attributes captured by the mobile device in order to obtain the final

decryption key. Then, the author proposed the third scheme, the low-complexity

MA-ABE scheme, which effectively offloads the high-cost computational and

communication workloads to Cloud based Semi Trusted Authority (STA). By

183
 Chapter 6. Context-Aware Attribute-Based Encryption Schemes

analysing the second scheme, the author found that the performance of mobile

device to complete all the processes may degrade together with the user experience.

High computation workloads should be offloaded to the Cloud. The STA which

works on behalf of the user is introduced, thus the performance of mobile devices

are much improved. The communication overheads and computation work are

reduced compared with other schemes. Meanwhile, the designed protocol

maintains the system security and data access at the same level as the previous

schemes.

184
7 Conclusions and Future Work

7.1 Summary and Conclusions

This thesis investigated security and privacy issues in the emerging mobile Cloud

paradigm. Traditional access control techniques developed for Cloud environment

do not support both the privacy and security issues in the mobile Cloud

environment. In mobile Cloud environment, due to the seamless interaction

between the user and smart devices, the user generates a large volume of personal

data. These data are eventually collected by the mobile Cloud service providers and

users do not have much control over their utilisation. In order to enhance the user

privacy together with security in the mobile Cloud environment, various access

control techniques were proposed in this thesis. The conventional attribute-based

access control and attribute-based encryption (ABE) schemes were modified in the

proposed schemes in order to support the access control in mobile Cloud

environments.

Mobile Cloud environment is a combination of Cloud computing technology

together with mobile Internet. It delivers tailored conventional Cloud services to

smart mobile devices. This enables the smart devices to offload computational,

communicational, and storage tasks to the Cloud servers. In mobile Cloud

environment, Cloud service providers collect very sensitive real-time user

information such as location, user behaviour and contact details. The existing

management systems such as .NET Passport, OpenID, Liberty Alliance, Higgins,

185
 Chapter 7. Conclusions and Future Work

and OAuth were solely developed to protect user‟s identities in conventional Cloud

environments. The author compared these identity management systems and

investigated the issues where such identity management systems are used in mobile

Cloud environments. Moreover, conventional access control mechanisms such as

DAC, ManAC, RBAC, and ABAC, privacy protection languages such as P3P,

EPAL, and XACML and ABE encryption techniques cannot be directly extended to

the mobile Cloud environment.

For mobile Cloud environment, three techniques were proposed to protect users‟

privacy and security of online data: (1) deploying an access control technique to

control the usage of user data in Chapter 5, (2) privacy-preserving technique to

protect the user data in Chapters 5 and 6, and (3) storing the online data in the

encrypted form in Chapter 6. Implementation of a user-centric attribute-based

accesscontrol model is also given for mobile Cloud environments. The proposed

techniques mutually authorize the user and the service provider (SP).

XACML-based access control protocol was proposed to enable mobile usersto

define access control policies for their online data. SPs need to satisfy the privacy

access policy in order to obtain users‟ data. An attribute authority (AA) is proposed

to maintain attributes of the users and SPs. This assures the status of a SP so that

user‟s sensitive data will not be revealed to a malicious third party. The available

real-time context-related attributes such as time, date, and location are used during

the authorization process. When a requester requests user‟s privacy data, the system

will pick required real-time attributes to verify the requesterduring the policy

evaluation process. The model is built on top of the XACML standard, which is one

of the popular privacy-preserving languages. he standard XACML architecture was

186
 Chapter 7. Conclusions and Future Work

extended by using attributes from the mobile users, SPs, and surrounding

environment so that the authorization can be done at runtime. Thus, the access

control of user‟s data in mobile Cloud environment is more secure than the

traditional access control schemes.

Using XACML requires the system to follow the standards of XACML. All the

message flows should be formatted into the XACML specification, which is a

drawback in the collaboration environment such as the mobile Cloud environment.

Different domains would define their own structures and protocols to set up an

access control system. For a mobile user of domain A, it would be infeasible to

request data from domain B if domain B could not understand the requested

message from domain A. Furthermore, policy evaluation would be a long-time

process if there are a large number of policies. In order to address these issues, the

author investigated the ABE schemes.

ABE scheme was modified to ensure the security of data together with privacy.

There were three different ABE schemes proposed in this thesis for data and

services access in the mobile Cloud environment. In the proposed schemes, the data

are stored in the encrypted formatso that data confidentiality is guaranteed. ABE

enables the data owner to define access policy during the encryption process. Due

to the features of mobile users and mobile platforms in mobile Cloud environment,

contextual attributes are available and can be used to strengthen the security and

data privacy. The first technique is a single authority context-aware ABE scheme

which incorporates contextual attributes. The data owner defines privacy access

policy during the encryption process and uploads the encrypted data. An AA

maintains a set of attributes, issues decryption keys and can monitor all the

187
 Chapter 7. Conclusions and Future Work

communications. After receiving the decryption keys from the AA, the contextual

attributes of the data requester are also collected by data requester‟s mobile device.

These attributes are combined with the decryption keys to recover the original data.

Single authority ABE scheme is impractical since the single authority has too

much control, such as manages users‟ attributes, issues decryption keys, and

monitor communication. All the data will be lost if it is compromised by an attacker.

Also single authority system cannotbe scaled due to the bottleneck problem if there

are a large number of users requesting keys. Meanwhile, in reality, different sets of

attributes are stored in different AAs. In order to incorporate this reality, a

context-aware multi-authority attribute-based encryption (MA-ABE) scheme was

proposed. More than one AA maintains different sets of attributes and issues

decryption keys. In order to protect user‟s identity from being tracked by each AA,

an anonymous key issuing protocol was proposed. Using this protocol, the data

requester‟s real identity cannot be revealed to each AA.Hence, an AA cannot track

the requester‟s data transactions. Due to the nature of the anonymous key issuing

protocol, as long as there are two honest authorities and other AAs are malicious;

the rest of the malicious AAs cannot combine the requester‟s data that they hold to

recover the original text. Thus, the collusion attacks are prevented. The proposed

scheme removed the central authority, which takes charge of issuing decryption

keys and monitoring all the activities. As a result, the central authority will not be

the bottleneck if the system is designed in a large scale. The loss of corruption of

central authority can be avoided using the proposed scheme.

In order to decrypt the ciphertext, the data requester must satisfy the access

policy. The requestor‟s mobile device capturesrequired contextual attributes and

188
 Chapter 7. Conclusions and Future Work

combines them with the decryption keys received from all the AAs. The data

requester has to contact all the AAs to get the decryption keys. The communication

overheads will be high if there are a large number of AAs. In mobile Cloud

environment， considering network conditions, the reliability of data network to

complete such communicationsis unpredictable.This can bea drawback.

Furthermore, final decryption key is computed based on the keys received from all

the AAs and attributes captured by the mobile devices, hence the computational

work is a burden. Those two limitations affect the performances of the mobile

Cloud services and lower the quality of services.

A low-complexity ABE scheme was proposed to address the issues that exist in the

context-aware MA-ABE scheme as a third technique. A Cloud-based Semi-Trusted

Authority (STA) is introduced, thus, the high-cost communications and

computationsare migrated from the mobile end to the Cloud end. This method

improves the performance at the mobile end. In this proposed algorithm, a user is

not required to make contact with the AAs; STA requests the keys on behalf of the

user. The user has a pre-shared secret with each AA, which is unknown to the STA.

When issuing decryption keys, each AA embeds the pre-shared secret into the

decryption keys.Therefore, STA cannot compute the final decryption key to

recover the message. The security analysis showed that the proposed

low-complexity technique prevents collusion attacks. The proposed techniques not

only secure the users‟ data, but also empower the mobile users to control the

privacy of their data.

189
 Chapter 7. Conclusions and Future Work

Finally, this thesis proposes the data access control techniques and ABE

schemes to protect users‟ privacy and sensitive data from unauthorized access.

In this part of the research work, an attributes-based access control model for

mobile Cloud environment is proposed to restrict access of users‟ online data. The

author deployed access control techniques to protect the user‟s data access.

XACML is used as the privacy access policy language, which is a flexible and

user-centric solution for mobile environment..

(1) The MA-ABE scheme for mobile Cloud environments is also

developed and implemented to protect user privacy and data

security from unauthorized access and fraud. As each of the AAs

maintains different sets of attributes for issuing decryption keys for

the use. The MA-ABE scheme removes the central authority and

improves the system performance by removing single point of

failure and providing end to end security

(2) The incorporation of context-related attributes compared with the

conventional ABE enables the data owner to define the privacy

access policies for their personal data at runtime. The mobile

devices of users will capture the dynamic contextual attributes to

satisfy the privacy access policies. It strengthens the privacy and

data protection in a more flexible, user-centric, and fine-grained

manner.

(3) In addition to this, an anonymisation key issuing protocol is

designed to protect the users‟ identity during data transactions. An

190
 Chapter 7. Conclusions and Future Work

AA cannot track a transaction and identify the user who initiates the

transaction. Furthermore, the AAs cannot combine the information

they hold to reveal the original text as long as there are two honest

authorities. The proposed scheme maintains the collusion resistance

against up to (N – 2) AAs.

7.2 Recommendations for Future Work

The following summaries some of the future research work that could be carried out

in this area based on our research findings.

 The use of context-related attributes can be further investigated.

With the development of the mobile operating systems, mobile

applications for detecting and analysing user‟s behaviours can

derive more valuable and accurate contextual attributes to define the

risk level of a user. The incorporation of such applications with

Cloud based services can be investigated to develop highly secure

solutions in mobile Cloud environments.

 Enable the data owner to define more complex access structures to

restrict the access to his online resources.

 Adding a key revocation mechanism. If a user no longer holds an

attribute, the relevant private key should be updated by the authority

in parallel. The user could not use the previous key to operate the

resource any longer.

191
 Chapter 7. Conclusions and Future Work

 Developing completely decentralized ABE scheme where new AA

can join and leave the mobile Cloud environment without

bootstrapping the whole system. This enables seamless interaction

between the user and SPs and will be ideal for multi-Cloud

environments. However, developing such algorithms under standard

security assumptions such as decisional bilinear Diffie-Hellman is

very challenging at the same time promising.

192
Bibligraphy:

[1] S. Weiguang and S. Xiaolong, "Review of Mobile cloud computing," in

Communication Software and Networks (ICCSN), 2011 IEEE 3rd
International Conference on, 2011, pp. 1-4.

[2] C. Ivan and R. Popa, "Cloud based Cross Platform Mobile Applications
Building and integrating cloud services with mobile client applications,"
Advances in Computer Science: an International Journal, vol. 3, pp. 69-77,
2014.

[3] S. Hui, L. Zhuohua, W. Jiafu, and Z. Keliang, "Security and privacy in

mobile cloud computing," in Wireless Communications and Mobile
Computing Conference (IWCMC), 2013 9th International, 2013, pp.
655-659.

[4] V. C. LEUNG, Y. WEN, M. CHEN, and C. RONG, "MOBILE CLOUD

COMPUTING," IEEE Wireless Communications, p. 13, 2013.

[5] A. N. Khan, M. Mat Kiah, S. U. Khan, and S. A. Madani, "Towards secure

mobile cloud computing: A survey," Future Generation Computer Systems,
vol. 29, pp. 1278-1299, 2013.

[6] H. T. Dinh, C. Lee, D. Niyato, and P. Wang, "A survey of mobile cloud
computing: architecture, applications, and approaches," Wireless
Communications and Mobile Computing, vol. 13, pp. 1587-1611, 2013.

[7] W. Jansen and T. Grance, "Guidelines on security and privacy in public

cloud computing," NIST special publication, vol. 800, p. 144, 2011.

[8] K. Sangani, "Rolling out the mobile future," Engineering & Technology,
vol. 7, pp. 80-81, 2012.

[9] S.-Z. Yang, "The marketing chain in the mobile Internet era," in Machine
Learning and Cybernetics (ICMLC), 2011 International Conference on,
2011, pp. 1058-1061.

[10] E. report. (Jan, 2013). Three Out of Four UK Mobile Users to Own
Smartphones by 2016. Available:
http://www.emarketer.com/Article/Three-of-Four-UK-Mobile-Users-Own-
Smartphones-by-2016/1009614

[11] W3C. (2014). Extensible Markup Language (XML). Available:

http://www.w3.org/XML/

193
 Bibliography

[12] D. Crockford. (2014). Introducing JSON. Available: http://json.org/

[13] M. Gudgin, M. Hadley, N. Mendelsohn, J.-J. Moreau, H. F. Nielsen, A.

Karmarkar, and Y. Lafon, "Simple object access protocol (SOAP) 1.2,"
World Wide Web Consortium, 2003.

[14] (Apr, 2013). Apple iOS. Available: http://www.apple.com/uk/ios/

[15] (Apr. 2013). Google Android. Available: http://www.android.com/

[16] Statista. (June 2014). Cumulative number of apps downloaded from the
Apple App Store from June 2008 to June 2014. Available:
http://www.statista.com/statistics/263794/number-of-downloads-from-the-
apple-app-store/

[17] Statista. (Sep. 2014). Number of available apps in the Apple App Store from
July 2008 to September 2014. Available:
http://www.statista.com/statistics/263795/number-of-available-apps-in-the
-apple-app-store/

[18] Statista. (July 2013). Cumulative number of apps downloaded from the
Google Play Android app store as of July 2013. Available:
http://www.statista.com/statistics/281106/number-of-android-app-downloa
ds-from-google-play/

[19] V. H. (24, July, 2013). Android's Google Play beats App Store with over 1
million apps, now officially largest. Available:
http://www.phonearena.com/news/Androids-Google-Play-beats-App-Store
-with-over-1-million-apps-now-officially-largest_id45680

[20] AppBrain. (2014). Number of Android applications. Available:

http://www.appbrain.com/stats/number-of-android-apps

[21] T. Soyata, H. Ba, W. Heinzelman, M. Kwon, and J. Shi, "Accelerating

mobile cloud computing: A survey," Communication Infrastructures for
Cloud Computing, 2013.

[22] H. T. Dinh, C. Lee, D. Niyato, and P. Wang, "A survey of mobile cloud
computing: architecture, applications, and approaches," Wireless
Communications and Mobile Computing, 2011.

[23] M. Almorsy, J. Grundy, and A. S. Ibrahim, "Collaboration-based cloud

computing security management framework," in Cloud Computing
(CLOUD), 2011 IEEE International Conference on, 2011, pp. 364-371.

[24] M. Armbrust, A. Fox, R. Griffith, A. D. Joseph, R. Katz, A. Konwinski, G.

Lee, D. Patterson, A. Rabkin, and I. Stoica, "A view of cloud computing,"
Communications of the ACM, vol. 53, pp. 50-58, 2010.

194
 Bibliography

[25] P. Mell and T. Grance, "The NIST definition of cloud computing (draft),"
NIST special publication, vol. 800, p. 7, 2011.

[26] L. Zhong, B. Wang, and H. Wei, "Cloud computing applied in the mobile
internet," in Computer Science & Education (ICCSE), 2012 7th
International Conference on, 2012, pp. 218-221.

[27] R. Jain, "Quality of experience," IEEE MultiMedia, vol. 11, pp. 96-95,
2004.

[28] Z. Wang and J. Crowcroft, "Quality-of-service routing for supporting

multimedia applications," Selected Areas in Communications, IEEE
Journal on, vol. 14, pp. 1228-1234, 1996.

[29] F. c. G. LLC. (2013). The Definition of Customer Quality of Experience.

Available:
https://www.findyourcloud.com/Articles/Quality-of-Experience/The-Defin
ition-of-Customer-Quality-of-Experience/

[30] Z. Peng and Y. Zheng, "A QoS-aware system for mobile cloud computing,"
in Cloud Computing and Intelligence Systems (CCIS), 2011 IEEE
International Conference on, 2011, pp. 518-522.

[31] F. c. G. LLC. (2013). How to Measure QX Evidence. Available:

http://www.findyourcloud.com/Articles/Research-2.0/How-to-Measure-Q
X-Evidence/

[32] (2013). Instagram. Available: http://instagram.com/

[33] (2013). Flickr. Available: http://www.flickr.com/

[34] (2013). Facebook. Available: https://www.facebook.com/

[35] Gartner. (April 2013). Gartner Says By 2016, 40 Percent of Mobile

Application Development Projects Will Leverage Cloud Mobile Back-End
Services. Available: http://www.gartner.com/newsroom/id/2463615

[36] L. Zhang, X. Ding, Z. Wan, M. Gu, and X.-Y. Li, "WiFace: a secure
geosocial networking system using WiFi-based multi-hop MANET," in
Proceedings of the 1st ACM Workshop on Mobile Cloud Computing &
Services: Social Networks and Beyond, 2010, p. 3.

[37] M. Pitkänen, T. Kärkkäinen, J. Ott, M. Conti, A. Passarella, S. Giordano, D.

Puccinelli, F. Legendre, S. Trifunovic, and K. Hummel, "SCAMPI: Service
platform for social aware mobile and pervasive computing," ACM
SIGCOMM Computer Communication Review, vol. 42, pp. 503-508, 2012.

[38] D. Recordon and D. Reed, "OpenID authentication 2.0-final," ed:

December, 2007.

195
 Bibliography

[39] Eclipse. (2011). Higgins 2.0. Available: http://eclipse.org/higgins/

[40] T. LLC. (2013). Tripadvisor. Available: http://www.tripadvisor.co.uk/

[41] O. K. a. S. Moritz. (2013). Carriers Sell Users' Tracking Data in $5.5

Billion Market. Available:
http://www.bloomberg.com/news/2013-06-06/carriers-sell-users-tracking-
data-in-5-5-billion-market.html

[42] G. Sloane. (2014). Foursquare Starts Selling All That Data. Available:
http://www.adweek.com/news/technology/foursquare-starts-selling-all-dat
a-158628

[43] A. TROIANOVSKI. (2013). Phone Firms Sell Data on Customers.

Available:
http://online.wsj.com/news/articles/SB10001424127887323463704578497
153556847658?mg=reno64-wsj&url=https%3A%2F%2Ffanyv88.com%3A443%2Fhttp%2Fonline.wsj.com%2
Farticle%2FSB10001424127887323463704578497153556847658.html

[44] (January 2014) Mobile Malware Infects Millions; LTE Spurs Growth.
Infosecurity. Available:
http://www.infosecurity-magazine.com/news/mobile-malware-infects-mill
ions-lte-spurs-growth/

[45] S. Bugiel, L. Davi, A. Dmitrienko, T. Fischer, and A.-R. Sadeghi,

"Xmandroid: A new android evolution to mitigate privilege escalation
attacks," Technische Universität Darmstadt, Technical Report TR-2011-04,
2011.

[46] S. Bugiel, L. Davi, A. Dmitrienko, T. Fischer, A.-R. Sadeghi, and B.

Shastry, "Towards Taming Privilege-Escalation Attacks on Android," in
NDSS, 2012.

[47] O. Publishing, OECD Guidelines on the Protection of Privacy and

Transborder Flows of Personal Data: Organisation for Economic
Co-operation and Development, 2002.

[48] U. S. D. o. H. H. Services. HIPAA. Available:

http://www.hhs.gov/ocr/privacy/

[49] E. Commission. (2013). Protection of personal data. Available:

http://ec.europa.eu/justice/data-protection/

[50] E. Commission, "The Data Protection Directive (Directive 95/46/EC),"

2013.

[51] S. S. Greene, Security Policies and Procedures: New Jersey: Pearson

Education, 2006.

196
 Bibliography

[52] G. Stoneburner, C. Hayden, and A. Feringa, "Engineering principles for

information technology security (a baseline for achieving security)," DTIC
Document2001.

[53] J. Jonsson and B. Kaliski, "Public-key cryptography standards (PKCS)# 1:

RSA cryptography specifications version 2.1," 2003.

[54] T. Dierks, "The transport layer security (TLS) protocol version 1.2," 2008.

[55] N.-F. Standard, "Announcing the Advanced Encryption Standard (AES),"

Federal Information Processing Standards Publication, vol. 197, 2001.

[56] P. Karn, W. A. Simpson, and P. Metzger, "The ESP triple DES transform,"
1995.

[57] K. Kaukonen and R. Thayer, "A stream cipher encryption algorithm

“arcfour”," The Internet Society, 1999.

[58] T. ElGamal, "A public key cryptosystem and a signature scheme based on
discrete logarithms," in Advances in Cryptology, 1985, pp. 10-18.

[59] R. L. Rivest, A. Shamir, and L. Adleman, "A method for obtaining digital
signatures and public-key cryptosystems," Communications of the ACM,
vol. 21, pp. 120-126, 1978.

[60] C. Furlani, "FIPS 186-3: Digital Signature Standard (DSS)," Online,

National Institute of Standards and Technology (NIST) Std.(June 2009),
2009.

[61] M. Bellare, R. Canetti, and H. Krawczyk, "Keying hash functions for

message authentication," in Advances in Cryptology—CRYPTO’96, 1996,
pp. 1-15.

[62] S. H. Standard, "FIPS Pub 180-1," National Institute of Standards and

Technology, vol. 17, p. 15, 1995.

[63] X. Wang, Y. L. Yin, and H. Yu, "Finding collisions in the full SHA-1," in
Advances in Cryptology–CRYPTO 2005, 2005, pp. 17-36.

[64] A. C. Weaver, "Secure Sockets Layer," Computer, vol. 39, pp. 88-90, 2006.

[65] K. Hickman and T. Elgamal, "The SSL protocol," Netscape

Communications Corp, vol. 501, 1995.

[66] P. Hallam-Baker and E. Maler, "Assertions and protocol for the oasis
security assertion markup language (saml)," OASIS XML-Based Security
Services Technical Committee, 2002.

197
 Bibliography

[67] S. Cantor, I. J. Kemp, N. R. Philpott, and E. Maler, "Assertions and

protocols for the oasis security assertion markup language," OASIS
Standard (March 2005), 2005.

[68] W. Redmond, "Microsoft Passport: Streamlining Commerce and

Communication on the Web," Microsoft News Center, 1999.

[69] M. Walker. (27 Oct 2008). Windows Live ID now supports OpenID Identity
Framework. Available:
http://blogs.msdn.com/b/mikewalker/archive/2008/10/27/windows-live-id-
will-support-openid-identity-framework.aspx

[70] D. Burt. (27 Aug 2009). Windows Live ID OpenID Status Update. Available:
http://blogs.technet.com/b/privacyimperative/archive/2009/08/28/windows
-live-id-openid-status-update.aspx

[71] S. S. Y. Shim, B. Geetanjali, and P. Vishnu, "Federated identity

management," Computer, vol. 38, pp. 120-122, 2005.

[72] H. S. Al-Sinani, W. A. Alrodhan, and C. J. Mitchell, "CardSpace-Liberty

integration for CardSpace users," in Proceedings of the 9th Symposium on
Identity and Trust on the Internet, 2010, pp. 12-25.

[73] S. Cantor, J. Hodges, J. Kemp, and P. Thompson, "Liberty ID-FF

Architecture Overview," Wason, Thomas (Herausgeber): Liberty Alliance
Project Version, vol. 1, 2005.

[74] J. Tourzan and Y. Koga, "Liberty id-wsf web services framework

overview," Liberty Alliance, 2004.

[75] S. Kellomäki and R. Lockhart, "Liberty ID-SIS Employee Profile Service

Specification," ed: Version, 1945.

[76] D. Recordon and D. Reed, "OpenID 2.0: a platform for user-centric identity
management," in Proceedings of the second ACM workshop on Digital
identity management, 2006, pp. 11-16.

[77] W. Huping, F. Chunxiao, Y. Shuai, Z. Junwei, and Z. Xiaoying, "A New

Secure OpenID Authentication Mechanism Using One-Time Password
(OTP)," in Wireless Communications, Networking and Mobile Computing
(WiCOM), 2011 7th International Conference on, 2011, pp. 1-4.

[78] O. Hyun-Kyung and J. Seung-Hun, "The Security Limitations of SSO in

OpenID," in Advanced Communication Technology, 2008. ICACT 2008.
10th International Conference on, 2008, pp. 1608-1611.

[79] D. Recordon and B. Fitzpatrick, "OpenID Authentication 1.1," Finalized

OpenID Specification, May, 2006.

198
 Bibliography

[80] G. Alpár, J.-H. Hoepman, and J. Siljee, "The identity crisis. security,
privacy and usability issues in identity management," arXiv preprint
arXiv:1101.0427, 2011.

[81] (2013). OAuth. Available: http://oauth.net/about/

[82] E. Hammer-Lahav, D. Recordon, and D. Hardt, "The OAuth 2.0

authorization protocol," Network Working Group Internet-Draft, 2011.

[83] H. Takabi, J. B. Joshi, and G.-J. Ahn, "Security and Privacy Challenges in
Cloud Computing Environments," IEEE Security & Privacy, vol. 8, pp.
24-31, 2010.

[84] Z. Peng, S. Hanlin, and Y. Zheng, "Building up Trusted Identity

Management in Mobile Heterogeneous Environment," in Trust, Security
and Privacy in Computing and Communications (TrustCom), 2011 IEEE
10th International Conference on, 2011, pp. 873-877.

[85] B. En-Nasry and M. D. E.-C. El Kettani, "Towards an Open Framework for

Mobile Digital Identity Management through Strong Authentication
Methods," in Secure and Trust Computing, Data Management, and
Applications, ed: Springer, 2011, pp. 56-63.

[86] V. Paruchuri and S. Chellappan, "Context Aware Identity Management

Using Smart Phones," in Broadband and Wireless Computing,
Communication and Applications (BWCCA), 2013 Eighth International
Conference on, 2013, pp. 184-190.

[87] M. Kim, H. Jeong, and E. Choi, "Context-aware Platform for User

Authentication in Cloud Database Computing," in International Conference
on Future Information Technology and Management Science &
Engineering Lecture Notes in Information Technology, 2012, pp. 170-176.

[88] H. Witte, C. Rathgeb, and C. Busch, "Context-Aware Mobile Biometric

Authentication based on Support Vector Machines," in Emerging Security
Technologies (EST), 2013 Fourth International Conference on, 2013, pp.
29-32.

[89] D. Huang, T. Xing, and H. Wu, "Mobile cloud computing service models: a
user-centric approach," IEEE Network, vol. 27, pp. 6-11, 2013.

[90] D. Boger, L. Barreto, J. Fraga, P. Urien, H. Aissaoui, A. Santos, and G.

Pujolle, "User-centric Identity Management based on secure elements," in
Computers and Communication (ISCC), 2014 IEEE Symposium on, 2014,
pp. 1-6.

[91] S. Sin-seok, K. Joon-Myung, H. Yoonseon, and J. W. K. Hong, "Context

management for user-centric context-aware services over pervasive

199
 Bibliography

networks," in Network Operations and Management Symposium

(APNOMS), 2012 14th Asia-Pacific, 2012, pp. 1-4.

[92] M. Bourimi, J. M. Haake, M. Heupel, B. Ueberschär, D. Kesdogan, and T.

Barth, "Enhancing privacy in mobile collaborative applications by enabling
end-user tailoring of the distributed architecture," International Journal for
Infonomics (IJI), vol. 3, pp. 563-572, 2011.

[93] F. Li, N. Clarke, M. Papadaki, and P. Dowland, "Behaviour profiling on

mobile devices," in Emerging Security Technologies (EST), 2010
International Conference on, 2010, pp. 77-82.

[94] L. Ki Jung and S. Il-Yeol, "Modeling and Analyzing User Behavior of

Privacy Management on Online Social Network: Research in Progress," in
Privacy, Security, Risk and Trust (PASSAT) and 2011 IEEE Third
Inernational Conference on Social Computing (SocialCom), 2011 IEEE
Third International Conference on, 2011, pp. 1344-1351.

[95] P. Paraskevopoulos, T. Dinh, Z. Dashdorj, T. Palpanas, and L. Serafini,

"Identification and characterization of human behavior patterns from
mobile phone data," in Intl. Conf. the Analysis of Mobile Phone Datasets
(NetMob 2013), Special Session on the Data for Development (D4D)
Challenge, 2013.

[96] N. Lathia, V. Pejovic, K. K. Rachuri, C. Mascolo, M. Musolesi, and P. J.

Rentfrow, "Smartphones for Large-Scale Behavior Change Interventions,"
IEEE Pervasive Computing, vol. 12, pp. 66-73, 2013.

[97] H. T. Tavani and J. H. Moor, "Privacy protection, control of information,

and privacy-enhancing technologies," SIGCAS Comput. Soc., vol. 31, pp.
6-11, 2001.

[98] A. RFC, "Internet Security Glossary," ed: Version, 2007.

[99] H. M. Levy, Capability-based computer systems vol. 12: Digital Press

Bedford, 1984.

[100] R. S. Sandhu and P. Samarati, "Access control: principle and practice,"

Communications Magazine, IEEE, vol. 32, pp. 40-48, 1994.

[101] C. S. Jordan, Guide to Understanding Discretionary Access Control in

Trusted Systems: DIANE Publishing, 1987.

[102] B. W. Lampson, "Protection," ACM SIGOPS Operating Systems Review,

vol. 8, pp. 18-24, 1974.

[103] R. N. Watson, "A decade of OS access-control extensibility,"

Communications of the ACM, vol. 56, pp. 52-63, 2013.

200
 Bibliography

[104] P. Samarati and S. C. de Vimercati, "Access control: Policies, models, and

mechanisms," in Foundations of Security Analysis and Design, ed: Springer,
2001, pp. 137-196.

[105] E. Bertino, C. Bettini, and P. Samarati, "A discretionary access control

model with temporal authorizations," in New Security Paradigms Workshop,
1994. Proceedings., 1994 ACM SIGSAC, 1994, pp. 102-107.

[106] K. Lehmann and F. Matthes, "Meta model based integration of role-based

and discretionary access control using path expressions," in E-Commerce
Technology, 2005. CEC 2005. Seventh IEEE International Conference on,
2005, pp. 443-446.

[107] L. Ninghui and M. V. Tripunitara, "On safety in discretionary access

control," in Security and Privacy, 2005 IEEE Symposium on, 2005, pp.
96-109.

[108] T. Thomas, "A mandatory access control mechanism for the Unix file
system," in Aerospace Computer Security Applications Conference, 1988.,
Fourth, 1988, pp. 173-177.

[109] W. Rjaibi and P. Bird, "A multi-purpose implementation of mandatory

access control in relational database management systems," in Proceedings
of the Thirtieth international conference on Very large data bases-Volume
30, 2004, pp. 1010-1020.

[110] L. LI, Y.-Z. HE, and D.-G. FENG, "A fine-grained mandatory access
control model for XML documents," Journal of software, vol. 15, pp.
1528-1537, 2004.

[111] L. Sung-Min, S. Sang-bum, J. Bokdeuk, and M. Sangdok, "A Multi-Layer

Mandatory Access Control Mechanism for Mobile Devices Based on
Virtualization," in Consumer Communications and Networking Conference,
2008. CCNC 2008. 5th IEEE, 2008, pp. 251-256.

[112] R. J. Robles, C. Min-kyu, Y. Sang-Soo, and K. Tai-hoon, "Application of

Role-Based Access Control for Web Environment," in Ubiquitous
Multimedia Computing, 2008. UMC '08. International Symposium on, 2008,
pp. 171-174.

[113] H. Lindqvist, "Mandatory access control," Master's Thesis in Computing

Science, Umea University, Department of Computing Science, SE-901, vol.
87, 2006.

[114] R. Watson, W. Morrison, C. Vance, and B. Feldman, "The TrustedBSD

MAC Framework: Extensible Kernel Access Control for FreeBSD 5.0," in
USENIX Annual Technical Conference, FREENIX Track, 2003, pp.
285-296.

201
 Bibliography

[115] D. F. Ferraiolo and D. R. Kuhn, "Role-based access controls," arXiv

preprint arXiv:0903.2171, 2009.

[116] R. S. Sandhu, E. J. Coyne, H. L. Feinstein, and C. E. Youman, "Role-based

access control models," Computer, vol. 29, pp. 38-47, 1996.

[117] A. C. O‟Connor and R. J. Loomis, "2010 Economic Analysis of Role-Based

Access Control," RTI International report for NIST, 2010.

[118] E. Bertino, P. A. Bonatti, and E. Ferrari, "TRBAC: A temporal role-based

access control model," ACM Transactions on Information and System
Security (TISSEC), vol. 4, pp. 191-233, 2001.

[119] E. Coyne and T. Weil, "An RBAC implementation and interoperability

standard: The INCITS cyber security 1.1 model," Security & Privacy, IEEE,
vol. 6, pp. 84-87, 2008.

[120] E. Bertino, B. Catania, M. L. Damiani, and P. Perlasca, "GEO-RBAC: a

spatially aware RBAC," presented at the Proceedings of the tenth ACM
symposium on Access control models and technologies, Stockholm,
Sweden, 2005.

[121] S. Oh and S. Park, "Task–role-based access control model," Information

Systems, vol. 28, pp. 533-562, 2003.

[122] M. Strembeck and G. Neumann, "An integrated approach to engineer and

enforce context constraints in RBAC environments," ACM Trans. Inf. Syst.
Secur., vol. 7, pp. 392-427, 2004.

[123] M. S. Kirkpatrick and E. Bertino, "Enforcing spatial constraints for mobile

rbac systems," in Proceedings of the 15th ACM symposium on Access
control models and technologies, 2010, pp. 99-108.

[124] B. Shafiq, J. B. Joshi, E. Bertino, and A. Ghafoor, "Secure interoperation in

a multidomain environment employing RBAC policies," Knowledge and
Data Engineering, IEEE Transactions on, vol. 17, pp. 1557-1577, 2005.

[125] J. B. D. Joshi, E. Bertino, U. Latif, and A. Ghafoor, "A generalized temporal

role-based access control model," Knowledge and Data Engineering, IEEE
Transactions on, vol. 17, pp. 4-23, 2005.

[126] D. Ferraiolo, R. Kuhn, and R. Sandhu, "RBAC Standard Rationale:

Comments on "A Critique of the ANSI Standard on Role-Based Access
Control"," Security & Privacy, IEEE, vol. 5, pp. 51-53, 2007.

[127] D. Kulkarni and A. Tripathi, "Context-aware role-based access control in

pervasive computing systems," presented at the Proceedings of the 13th
ACM symposium on Access control models and technologies, Estes Park,
CO, USA, 2008.

202
 Bibliography

[128] E. Yuan and J. Tong, "Attributed based access control (ABAC) for web
services," in Web Services, 2005. ICWS 2005. Proceedings. 2005 IEEE
International Conference on, 2005.

[129] E. Coyne and T. R. Weil, "ABAC and RBAC: Scalable, Flexible, and
Auditable Access Management," IT Professional, vol. 15, pp. 14-16, 2013.

[130] S. Hai-bo and H. Fan, "An Attribute-Based Access Control Model for Web
Services," in Parallel and Distributed Computing, Applications and
Technologies, 2006. PDCAT '06. Seventh International Conference on,
2006, pp. 74-79.

[131] B. Lang, I. Foster, F. Siebenlist, R. Ananthakrishnan, and T. Freeman, "A

Flexible Attribute Based Access Control Method for Grid Computing,"
Journal of Grid Computing, vol. 7, pp. 169-180, 2009/06/01 2009.

[132] R. Bobba, O. Fatemieh, F. Khan, C. A. Gunter, and H. Khurana, "Using

Attribute-Based Access Control to Enable Attribute-Based Messaging," in
Computer Security Applications Conference, 2006. ACSAC '06. 22nd
Annual, 2006, pp. 403-413.

[133] L. Wang and B. Wang, "Attribute-Based Access Control Model for Web
Services in Multi-Domain Environment," in Management and Service
Science (MASS), 2010 International Conference on, 2010, pp. 1-4.

[134] S. BEJI, Y. JAMMOUSSI, and N. EL KADHI, "Towards

context-awareness security for mobile applications."

[135] J. Li, I. Ari, J. Jain, A. H. Karp, and M. Dekhil, "Mobile in-store

personalized services," in Web Services, 2009. ICWS 2009. IEEE
International Conference on, 2009, pp. 727-734.

[136] H. Ould-Slimane, M. Bande, and H. Boucheneb, "WiseShare: A

collaborative environment for knowledge sharing governed by ABAC
policies," in Collaborative Computing: Networking, Applications and
Worksharing (CollaborateCom), 2012 8th International Conference on,
2012, pp. 21-29.

[137] H. Shen and Y. Cheng, "A Context-Aware Semantic-Based Access Control

Model for Mobile Web Services," in Advanced Research on Computer
Science and Information Engineering. vol. 153, G. Shen and X. Huang, Eds.,
ed: Springer Berlin Heidelberg, 2011, pp. 132-139.

[138] P. Kumaraguru, L. Cranor, J. Lobo, and S. Calo, "A survey of privacy

policy languages," in Workshop on Usable IT Security Management (USM
07): Proceedings of the 3rd Symposium on Usable Privacy and Security,
ACM, 2007.

203
 Bibliography

[139] K. Bohrer and B. Holland, "Customer profile exchange (cpexchange)

specification," ed, 2000.

[140] M. Schunter and C. Powers, "The enterprise privacy authorization language

(epal 1.1)," W3C Working Group, 2003.

[141] S. Godik, T. Moses, A. Anderson, B. Parducci, C. Adams, D. Flinn, G.

Brose, H. Lockhart, K. Beznosov, and M. Kudo, "extensible access control
markup language (xacml) version 1.0," ed, 2003.

[142] J. Reagle and R. Wenning, "P3P and Privacy on the Web FAQ," The World
Wide Web Consortium, http://www. w3. org/P3P/P3FAQ. html, vol. 12,
1997.

[143] P. C. K. Hung, E. Ferrari, and B. Carminati, "Towards standardized Web

services privacy technologies," in Web Services, 2004. Proceedings. IEEE
International Conference on, 2004, pp. 174-181.

[144] L. Cranor, M. Langheinrich, and M. Marchiori, "A P3P preference

exchange language 1.0 (APPEL1. 0)," W3C working draft, vol. 15, 2002.

[145] P. P. Privacy, "An Assessment of P3P and Internet Privacy," ed: Electronic
Privacy Information Center, 2000.

[146] P. A. S. H. G. Karjoth and C. P. M. S. E. Privacy, "Authorization Language

(EPAL 1.1) Oct. 1, 2003 IBM Research mts at zurich. ibm. com," Source:
http://www. zurich. ibm.
com/security/enterprise-privacy/epal/Specification.

[147] A. Barth, A. Datta, J. C. Mitchell, and H. Nissenbaum, "Privacy and

contextual integrity: framework and applications," in Security and Privacy,
2006 IEEE Symposium on, 2006, pp. 15 pp.-198.

[148] XAMCL and O. S. S. T. Committee, "extendible Access Control Markup

Language (xacml) committee specification 2.0," ed: Feb, 2005.

[149] C. A. Ardagna, S. D. C. d. Vimercati, S. Paraboschi, E. Pedrini, and P.

Samarati, "An XACML-based privacy-centered access control system,"
presented at the Proceedings of the first ACM workshop on Information
security governance, Chicago, Illinois, USA, 2009.

[150] A. Anderson, "A comparison of two privacy policy languages: EPAL and
XACML," Sun Microsystems, Inc.2005.

[151] OASIS. (2013). OASIS eXtensible Access Control Markup Language

(XACML) TC. Available:
https://www.oasis-open.org/committees/tc_home.php?wg_abbrev=xacml

204
 Bibliography

[152] A. Anderson and B. Devaraj, "XACML-Based Web Services Policy

Constraint Language (WS-PolicyConstraints)," Working Draft, vol. 6, p. 24,
2005.

[153] V. S. Y. Cheng, P. C. K. Hung, and D. K. W. Chiu, "Enabling Web Services

Policy Negotiation with Privacy preserved using XACML," in System
Sciences, 2007. HICSS 2007. 40th Annual Hawaii International Conference
on, 2007, pp. 33-33.

[154] A. Anderson and H. Lockhart, "SAML 2.0 profile of XACML," OASIS,

September, vol. 51, 2004.

[155] C. A. Ardagna, S. De Capitani di Vimercati, G. Neven, S. Paraboschi, F. S.

Preiss, P. Samarati, and M. Verdicchio, "Enabling Privacy-preserving
Credential-based Access Control with XACML and SAML," in Computer
and Information Technology (CIT), 2010 IEEE 10th International
Conference on, 2010, pp. 1090-1095.

[156] W. Hommel, "Using XACML for Privacy Control in SAML-Based Identity

Federations," in Communications and Multimedia Security. vol. 3677, J.
Dittmann, S. Katzenbeisser, and A. Uhl, Eds., ed: Springer Berlin
Heidelberg, 2005, pp. 160-169.

[157] N. Ulltveit-Moe and V. Oleshchuk, "Mobile Security with Location-Aware

Role-Based Access Control," in Security and Privacy in Mobile Information
and Communication Systems. vol. 94, R. Prasad, K. Farkas, A. Schmidt, A.
Lioy, G. Russello, and F. Luccio, Eds., ed: Springer Berlin Heidelberg,
2012, pp. 172-183.

[158] Q. Xuebing and C. Adams, "XACML-Based Policy-Driven Access Control

for Mobile Environments," in Electrical and Computer Engineering, 2006.
CCECE '06. Canadian Conference on, 2006, pp. 643-646.

[159] V. Goyal, O. Pandey, A. Sahai, and B. Waters, "Attribute-based encryption

for fine-grained access control of encrypted data," in Proceedings of the
13th ACM conference on Computer and communications security, 2006, pp.
89-98.

[160] A. Sahai and B. Waters, "Fuzzy identity-based encryption," in Advances in

Cryptology–EUROCRYPT 2005, ed: Springer, 2005, pp. 457-473.

[161] A. Shamir, "How to share a secret," Communications of the ACM, vol. 22,
pp. 612-613, 1979.

[162] J. Benaloh and J. Leichter, "Generalized secret sharing and monotone

functions," in Advances in Cryptology—CRYPTO’88, 1990, pp. 27-35.

205
 Bibliography

[163] J. Bethencourt, A. Sahai, and B. Waters, "Ciphertext-policy attribute-based

encryption," in Security and Privacy, 2007. SP'07. IEEE Symposium on,
2007, pp. 321-334.

[164] A. Lewko and B. Waters, "Decentralizing attribute-based encryption," in

Advances in Cryptology–EUROCRYPT 2011, ed: Springer, 2011, pp.
568-588.

[165] M. Chase, "Multi-authority attribute based encryption," in Theory of

Cryptography, ed: Springer, 2007, pp. 515-534.

[166] M. Chase and S. S. Chow, "Improving privacy and security in

multi-authority attribute-based encryption," in Proceedings of the 16th
ACM conference on Computer and communications security, 2009, pp.
121-130.

[167] K. Yang and X. Jia, "Attributed-Based Access Control for Multi-authority

Systems in Cloud Storage," in Distributed Computing Systems (ICDCS),
2012 IEEE 32nd International Conference on, 2012, pp. 536-545.

[168] M. Li, S. Yu, Y. Zheng, K. Ren, and W. Lou, "Scalable and secure sharing
of personal health records in cloud computing using attribute-based
encryption," 2013.

[169] J. A. Akinyele, M. W. Pagano, M. D. Green, C. U. Lehmann, Z. N. J.

Peterson, and A. D. Rubin, "Securing electronic medical records using
attribute-based encryption on mobile devices," presented at the Proceedings
of the 1st ACM workshop on Security and privacy in smartphones and
mobile devices, Chicago, Illinois, USA, 2011.

[170] H.-C. Liao and Y.-H. Chao, "A new data encryption algorithm based on the
location of mobile users," Information Technology Journal, vol. 7, pp.
63-69, 2008.

[171] N. Chen, M. Gerla, D. Huang, and X. Hong, "Secure, selective group

broadcast in vehicular networks using dynamic attribute based encryption,"
in Ad Hoc Networking Workshop (Med-Hoc-Net), 2010 The 9th IFIP
Annual Mediterranean, 2010, pp. 1-8.

[172] X. Zhiqian and K. M. Martin, "Dynamic User Revocation and Key

Refreshing for Attribute-Based Encryption in Cloud Storage," in Trust,
Security and Privacy in Computing and Communications (TrustCom), 2012
IEEE 11th International Conference on, 2012, pp. 844-849.

[173] T. Priebe, W. Dobmeier, and N. Kamprath, "Supporting attribute-based

access control with ontologies," in Availability, Reliability and Security,
2006. ARES 2006. The First International Conference on, 2006, p. 8 pp.

206
 Bibliography

[174] R. Dhamija and L. Dusseault, "The seven flaws of identity management:

Usability and security challenges," Security & Privacy, IEEE, vol. 6, pp.
24-29, 2008.

[175] E. McCallister, Guide to protecting the confidentiality of personally

identifiable information: DIANE Publishing, 2010.

[176] M. Georgiev, S. Iyengar, S. Jana, R. Anubhai, D. Boneh, and V. Shmatikov,

"The most dangerous code in the world: validating SSL certificates in
non-browser software," in Proceedings of the 2012 ACM conference on
Computer and communications security, 2012, pp. 38-49.

[177] H. Jhe-Yi, S. Chien-Cheng, L. Wei-Hsiang, and C. C. Ho, "Android-based

mobile payment service protected by 3-factor authentication and virtual
private ad hoc networking," in Computing, Communications and
Applications Conference (ComComAp), 2012, 2012, pp. 111-116.

[178] C. Cremers, "The Scyther tool: Automatic verification of security

protocols," ed, 2009.

[179] C. J. F. Cremers, "Scyther: Semantics and verification of security

protocols," vol. 68, ed, 2006.

[180] C. J. Cremers, "The Scyther Tool: Verification, falsification, and analysis of

security protocols," in Computer Aided Verification, 2008, pp. 414-418.

[181] Oracle. (2013). MySQL :: The world's most popular open source database.
Available: http://www.mysql.com/

[182] S. XACML, "Sun‟s XACML Implementation Programmer‟s Guide," ed,

2007.

[183] A. S. Foundation. (2013). Apache Tomcat Available:

http://tomcat.apache.org/

[184] T. E. Foundation. (2013). Eclipse - The Eclipse Foundation open source

community website. Available: http://www.eclipse.org/

[185] N. Community. (2013). Netbeans IDE. Available: https://netbeans.org/

[186] G. Karjoth and A. Schade, "SERIALIZATION OF XACML POLICIES,"

ed: Google Patents, 2009.

[187] N. Ulltveit-Moe and V. Oleshchuk, "Mobile Security with Location-Aware

Role-Based Access Control," in Security and Privacy in Mobile Information
and Communication Systems, ed: Springer, 2012, pp. 172-183.

[188] F. Li, N. Clarke, M. Papadaki, and P. Dowland, "Misuse detection for

mobile devices using behaviour profiling," International Journal of Cyber
Warfare and Terrorism (IJCWT), vol. 1, pp. 41-53, 2011.

207
 Bibliography

[189] M. Miettinen, P. Halonen, and K. Hatonen, "Host-based intrusion detection

for advanced mobile devices," in Advanced Information Networking and
Applications, 2006. AINA 2006. 20th International Conference on, 2006, pp.
72-76.

[190] N. Eagle and A. Pentland, "Reality mining: sensing complex social

systems," Personal and ubiquitous computing, vol. 10, pp. 255-268, 2006.

[191] P. A. Bernstein, V. Hadzilacos, and N. Goodman, Concurrency control and

recovery in database systems vol. 370: Addison-wesley New York, 1987.

[192] G. Weikum and G. Vossen, Transactional information systems: theory,

algorithms, and the practice of concurrency control and recovery: Elsevier,
2001.

[193] A. Sahai and B. Waters, "Fuzzy Identity-Based Encryption," in Advances in

Cryptology – EUROCRYPT 2005. vol. 3494, R. Cramer, Ed., ed: Springer
Berlin Heidelberg, 2005, pp. 457-473.

[194] Samsung. (2013). Samsung KNOX-Solutions-Security | Samsung. Available:

http://www.samsung.com/global/business/mobile/solution/security/samsun
g-knox

[195] Blackberry. (2013). Black Enterprise Server - BES - US. Available:

http://us.blackberry.com/business/software/bes/overview.html

[196] A. D. Caro. (2012). The Java Pairing Based Cryptography Library (jPBC).
Available: http://gas.dia.unisa.it/projects/jpbc/index.html

[197] Z. Sanaei, S. Abolfazli, A. Gani, and M. Shiraz, "SAMI: Service-based

arbitrated multi-tier infrastructure for Mobile Cloud Computing," in
Communications in China Workshops (ICCC), 2012 1st IEEE International
Conference on, 2012, pp. 14-19.

208
Appendix

XACML Language

XACML Policy
The following polices defines that any user who holds an email address
from the domain of “city.ac.uk” have the access to read resources at
http://www.city.ac.uk.

<?xml version="1.0" encoding="UTF-8"?>

<Policy
xmlns="urn:oasis:names:tc:xacml:2.0:policy:schema:os"
xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
PolicyId="urn:oasis:names:tc:xacml:2.0:conformance-test:IIC082:policy"
RuleCombiningAlgId="urn:oasis:names:tc:xacml:1.0:rule-combining-
algorithm:permit-overrides">
<Rule
RuleId="urn:oasis:names:tc:xacml:2.0:conformance-test:IIC082:rule"
Effect="Permit">
<Description>
Anyone from city.ac.uk can perform any action on any
resource.
</Description>
<Target>
<Subject>
<SubjectMatchMatchId="urn:oasis:names:tc:xacml:1.0:function:rf
c822Namematch">
<AttributeValueDataType="http://www.w3.org/2001/XMLSchema
#string">
city.ac.uk</AttributeValue>
<SubjectAttributeDesignator
SubjectCategory="urn:oasis:names:tc:xacml:1.0:subject-

209
 Bibliography

category:access-subject"
AttributeId="urn:oasis:names:tc:xacml:1.0:
subject:subject-id"
DataType="urn:oasis:names:tc:xacml:1.0:datatype:
rfc822Name"/>
</SubjectMatch>
</Subject>
<Resources>
<Resource>
<ResourceMatch
MatchId="urn:oasis:names:tc:xacml:1.0:
function:anyURI-equal">
<AttributeValue
DataType="http://www.w3.org/2001/
XMLSchema#anyURI">
http://www.city.ac.uk</AttributeValue>
<ResourceAttributeDesignator
AttributeId="urn:oasis:names:tc:xacml:1.0:
resource:resource-id"
DataType="http://www.w3.org/2001/
XMLSchema#anyURI"/>
</ResourceMatch>
</Resource>
</Resources>
<Actions>
<Action>
<ActionMatch
MatchId="urn:oasis:names:tc:xacml:1.0:
function:string-equal">
<AttributeValue
DataType="http://www.w3.org/2001/XMLSchema#
string">

210
 Bibliography

Read</AttributeValue>
<ActionAttributeDesignator
AttributeId="urn:oasis:names:tc:xacml:1.0:
action:action-id"
DataType="http://www.w3.org/2001/
XMLSchema#string"/>
</ActionMatch>
</Action>
</Actions>
</Target>
</Rule>
</Policy>

211
 Bibliography

XACML Request
A user who holds an email account “[email protected]” requests the
resources of School of Engineering and Mathematical Science at
http://www.city.ac.uk.

<?xml version="1.0" encoding="UTF-8"?>

<Request
xmlns="urn:oasis:names:tc:xacml:2.0:context:schema:os"
xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
xsi:schemaLocation="urn:oasis:names:tc:xacml:2.0:
context:schema:os
access_control-xacml-2.0-context-schema-os.xsd">
<Subject>
<Attribute
AttributeId="urn:oasis:names:tc:xacml:1.0:
subject:subject-id"
DataType="urn:oasis:names:tc:xacml:1.0:
data-type:rfc822Name">
<AttributeValue>[email protected]</AttributeValue>
</Attribute>
<AttributeAttributeId="urn:oasis:names:tc:xacml:2.0:
conformance-test:age"
DataType="http://www.w3.org/2001/XMLSchema#integer">
<AttributeValue>45</AttributeValue>
</Attribute>
</Subject>
<Resource>
<AttributeAttributeId="urn:oasis:names:tc:xacml:1.0:
resource:resource-id"
DataType="http://www.w3.org/2001/XMLSchema#anyURI">
<AttributeValue>http://www.city.ac.uk</AttributeValue>
</Attribute>
</Resource>

212
 Bibliography

<Action>
<AttributeAttributeId="urn:oasis:names:tc:xacml:1.0:
action:action-id"
DataType="http://www.w3.org/2001/XMLSchema#string">
<AttributeValue>read</AttributeValue>
</Attribute>
</Action>
</Request>

XACML Response
Regarding to the policy example and response context example, the user has
an email account “[email protected]” can obtain the data consent on the
resources at “http://www.city.ac.uk/”. The policy evaluation decision should be
Permit. The response context is listed as follows.

<Response>
<Result ResourceId="http://www.city.ac.uk/">
<Decision>Permit</Decision>
<Status>
<StatusCode Value="urn:oasis:names:tc:xacml:1.0:status:ok"/>
</Status>
</Result>
</Response>

213