[VOLUME 6 I ISSUE 1 I JAN.

– MARCH 2019] e ISSN 2348 –1269, Print ISSN 2349-5138

http://ijrar.com/ Cosmos Impact Factor 4.236

Design and Implementation of Secure File Storage using Distributed

Cloud Mechanism

Manoj V. Bramhe1 & Dr. Milind V. Sarode2 & Dr. Meenakshi S. Arya3
1Phd Research Scholar, Department of Computer Science & Engineering G.H. Raisoni College of
Engineering, Nagpur.
2Professor and Head, Department of Computer Engineering , Government Polytechnic,

Yeotmal.
3Associate Prof. & Head, Department of Computer. Science & Engineering, G.H. Raisoni College of

Engineering, Nagpur.

Received: January 09, 2019 Accepted: February 09, 2019

ABSTRACT: Cloud computing and the cloud based services like storage and application are the most
demanding resources over the internet from few years. With better network or internet connectivity these
become very easy tools for remotely managing the information and the files. As the system introduced it comes
along with multiple issues and threats also. Even after having better network security mechanism, data
protection algorithm, information privacy system service provider can’t ensure full proof solution for
information security. Let’s ignore the security about the network transfer and other part; finally file resides at
particular storage with company ownership which may be called storage service provider and the company
employee with higher authorities having always access to customer’s files. Here the proposed solution deals
with the malicious user attack security and prevent file access by utilizing the multiple cloud environment for
storing files.

Key Words: Cloud Computing, DFS, API

I. INTRODUCTION multiple parts at user application layer which will

Cloud Computing is popular model for be uploaded securely after applying encryption in
distributed data storage due to its various multiple clouds. Only authorized user or owner of
advantages like on demand network access with the files can download data with proper user
minimal cost as per user's usage but it faces authentication .
various security challenges as whole data is
available to cloud service provider. Many II.RELATED WORK
solutions were proposed for maintaining security Many researchers have proposed various
and privacy in cloud like encryption, replication, solutions to cloud storage but most of them are
VM isolation, etc [3] but most of which have focus related to single cloud provider where even
on single cloud environment where whole though cost of storage is less but has security
information is stored on single public cloud which concerns and storage failure issues due to vendor
creates security concerns as system administrator -lock-in problems. Some solutions pertain to
has full access to data hence malicious multi-cloud storage system as mentioned in [2],
administrator can use data for malevolent [6],[8],[9] where data is fragmented into multiple
purpose, vendor-lock-in problem has loss of full chunks and stored on existing public / private
information in case service provider stops his cloud infrastructure. Authors in [4] proposed four
services, loss of data integrity due to modification type of architecture for maintaining security of
of data by unauthorized persons [7]. data in cloud. Most of the papers like [11], [14],
All these issues with single cloud provider can be [16] have used first architecture of replication of
removed by moving to multi-cloud services application but our system uses combination of
wheredata is distributed partially among various those architectures by not only replicating data
clouds hence no entity will ever get complete set but dividing it into multiple parts. Tahoe-LAFS
of data a time. Distributed file system (DFS) are scheme was discussed in [7] where open source
used for managing files from multiple hosts. distributed file system Tahoe is used instead of
Our proposed solution is to provide other DFS for reliability purpose. System uses
secured and reliable storage service for multi encryption for confidentiality purpose and
cloud environment. It will fragment user files into achieves recovery of data using secret sharing
scheme. In [8] cloud storage service model for
Research Paper IJRAR- International Journal of Research and Analytical Reviews 𝗂379
[ VOLUME 6 I ISSUE 1 I JAN.– MARCH 2019] E ISSN 2348 –1269, PRINT ISSN 2349-5138
inter and intra cloud is proposed at Iaas level. and it’s API set to access the file storage by
Different data chunks of file are stored in various providing proper credentials to the connect to
VMs of single or multi cloud. User can store and multiple cloud storage. DFS methods or API set
retrieve data from multiple cloud .System uses are designed for three main modules like, A)
user authentication followed by file splitting / file Encryption and decryption module B)
retrieval by cloud manager interface and then it Construction and deconstruction module C) File
will be handover to multiple clouds/users. System transfer module.
supports both inter and intra cloud operations but A) Encryption and decryption module:
security is not enough as encryption methodology this module is mainly designed to utilize the
is not used. Some multi-cloud storage approaches encryption and decryption methods in order to
as discussed in [16] and [17] are similar to our encrypt the files or the file part to protect it from
system but they had worked on integrity, unauthorized read operations and later on
availability and vendor lock-in problems and not decrypt the file to allow user to read the file as it
paid attention to all the security parameters like was originally. This module utilizes the Advance
confidentiality . Our proposed system handles all Encryption Standard (AES) algorithm to perform
major parameters of security of maintaining the encryption and decryption task.
confidentiality, availability and integrity of data.
III. PROPOSED SYSTEM
Looking toward the limitations, described in the
literature review, it is necessary to have a novel
universal procedure to overcome those
limitations. As almost every big organization
having web identity and runs on hosting or cloud
environment, the organizational information
becomes important aspect. Even after cloud
provides industry a wide range of security and
system free from malicious software virus attacks,
still the data is not safe from malicious users Figure 1: Proposed System Architecture
having administrative rights. One having super
user rights can access the data from cloud storage B)Construction and deconstruction modules: As
with wrong intensions. After going through the proposed system describes its approach to
different solutions to reduce malicious user attack protect the file by dividing the file in multiple
we found keeping the data out of user reach will parts and store it at different locations or cloud
make data more secure than by any other way, but storage. This module is used for fragmentation of
we can’t forget the fact we have to keep the data original file into multiple parts as displayed in
somewhere. Hence the proposed system is figure 2 . First step system does is to separate
designed by keeping these scenarios in 100/3th part of the file and saves this 1/3rd part
consideration where we will distribute and secure as 1st part in local server. Next is to again split
the data at different location in order to hide remaining 66.67% part in to 2 equal parts and
original data directly from user. Here system will store it in two different files in public clouds, This
take smart decision on the basis of user request will work out in following ways,
and split the data after successful encryption in to Step 1: Collect the first part of the file in temp
different blocks and store it on the multiple cloud buffer,
storage, now malicious user can access the data Temp_Buff_1=MemStream.Take(100/3)
but of no use. Another point of consideration is the Step 2: Write the Temp bufferin first part of the
safety of data in case of critical condition. The file,
proposed system is less with capability to re FL.write(File_Chunk_A,Temp_Buff_1)
calculate the data from rest of the storage if any Step 3: Collect remaining part of file for further
one of the storage gets failure. separation,
Figure 1 describes the proposed system Temp_Buff_2=MemStream.Skip(100/3)
layered architecture along with applicable Step 4: Collect the second part of file for further
methods and the DFS structures. Overall system separation,
designed to have independent development Temp_Buff_1=MemStream.Take(100/2)
platform to design and develop multiple cloud Step 5: Write the Temp buffer in second part of
based file storage mechanism, where application the file,
will only responsible for utilizing the DFS system FL.write(File_Chunk_B,Temp_Buff_1)
𝗂380 IJRAR- International Journal of Research and Analytical Reviews Research Paper
[VOLUME 6 I ISSUE 1 I JAN. – MARCH 2019] e ISSN 2348 –1269, Print ISSN 2349-5138
http://ijrar.com/ Cosmos Impact Factor 4.236
Step 6: Collect remaining part of file as a last or 3rd
part of the file in buffer, All the modules defined are independent of
Temp_Buff_1=MemStream.Skip(100/2) another module as every module has its unique
Step 7: Write the Temp buffer in third part of the functionality and no integration required from
file, other modules but system can’t deny the fact that
FL.write(File_Chunk_C,Temp_Buff_1) these needs to construct in proper sequence to get
desired result. Below example explain process in
detail,
K=KeyGen(Size,Type);
FL=Enc(OriginalFile,K);
P[]=Split(FL,3);
Where, P[0]: First part P[1]: Second Part P[2]:
Third part
Finally,
Pt=Connect(IP_1);
SendFile(P[0],Pt);
Pt=Connect(IP_2);
SendFile(P[2],Pt);
Pt=Connect(IP_3);
SendFile(P[3],Pt);
This API sets are designed to allow user configure
Figure 2:File Split Technique the connection parameters of multiple cloud file
storage using FTP protocol mainly.
Above mentioned techniques get performed in
simple steps yet gives accuracy and stable V. EXPERIMENTAL RESULTS
performance almost every time and this allow As the system deals with multiple level and
system easy merging of file by simply copying all system to execute the defined process calculations
the parts sequentially in separate file in original for execution time defers time to time and file to
state. file. First file is uploaded to application server
C)File transfer Module: This module is hence resources like CPU, Hard disk and network,
mainly deal with the transporting file over TCP/IP comes in to the picture. Every resource has its
network using different protocols or services like own response time and instruction processing
FTP, Azure, and AWS. Even after this is the main capacity. With the help of figure 4, system
based of the system. describes its multiple levels of system process
time or the load.
IV. IMPLEMENTATION
Main idea behind the system is to implement
complete system as API or SDK set so that it can
be reused in multiple applications hence final
implementation will come in the form of
independent reusable packages like *.DLL in
windows architecture. Every sub class or sub
package will have different methods as per
defined in DFS structures and methods.

Figure 4: Time Progression Graph

Level 1: Connection time; in order to upload the
file over network server first system need to
establish the network communications with the
server. Connection time differs over local network
and internet speed and even in internet scenario
connection state like sleep or active take another
few seconds to wake up the network connection.

Figure 3: System API Set Model

Research Paper IJRAR- International Journal of Research and Analytical Reviews 𝗂381
[ VOLUME 6 I ISSUE 1 I JAN.– MARCH 2019] E ISSN 2348 –1269, PRINT ISSN 2349-5138
Level2: Handshake time; once the network single cloud storage system.Byremoving the file
communication is ready, every application ownership information mechanism system can
protocols have its own overheads which may prevent storage from malicious user also.
include authentication using credentials and this
takes another few seconds to start actual data VII. REFERENCES
transfer methods. 1. Zhifeng Xiao and Yang Xiao, “Security and
Privacy in Cloud Computing”, IEEE
Level 3: Transfer time; this is the main time Communications Surveys & Tutorials, March
required for actual file or data transfer takes place 2012
in the network. It may vary and totally depend on 2. MohammedA.AlZain, Eric Pardede, Ben Soh,
the size of the file currently queued for transfer. James A. Thom, “Cloud Computing Security:
From Single to Multi-Clouds”, IEEE 45th
Hawaii International Conference on System
Sciences, 2012
3. Singhal M., Chandrasekhar S., TingjianGe.,
Sandhu R., Krishnan R., Gail-JoonAhn., Bertino
E., “Collaboration in Multicloud Computing
Environments: Framework and Security
Issues”, IEEE computer society journal, Vol. 46,
Issue 2, pp. 76-84, Feb 2013
4. Bohli J., Gruschka N., Jensen M., Lo Iacono L.,
Marnau N, “Security and Privacy Enhancing
Multi-Cloud Architectures,” IEEE Transaction
on Dependable and secure computing, VolPP,
Issue 99, 2013
5. Tran Doan Thanh, Subaji Mohan, EunmiChoil,
Figure 5: Transfer Time Analyses SangBum Kim, Pilsung Kim “A Taxonomy and
Level 4: Network load; even if the transfer start Survey on Distributed File Systems,” IEEE
over the network, it is a continuous process Fourth International Conference on
executes till the end of the file transfer, this may Networked Computing and Advanced
depends on the network load by other Information Management, 2008
6. Su Chen, Yi Chen, Hai Jiang, Laurence T Yang,
applications running in the network over same
Kuan-Ching Li, “ A secure distributed file
interface. system based on revised Blakely’s secret
Figure 5 shows transfer time analysis for sharing scheme,” 11 th IEEE international
transferring the data over the network. Different conference on trust, security and privacy in
time periods are required for transferring the load computing and communications, 2012
depending upon various network parameters 7. Fan-Hsun Tseng, Chi-Yuan Chen, Li-Der Chou,
discussed in above section. Thus our total time Han-Chieh Chao, “Implement a reliable and
required to process the data is summation of secure cloud distributed file system,” IEEE
connection time, handshake time, transfer time international symposium on intelligent signal
processing and communication systems,
and network load time. All the above parameters
November 2012
affect the processing of data in distributed 8. ShushantShrivastava, Vikas Gupta, Rajesh
environment. Yadav, Krishna Kant, “Enhanced Distributed
storage on the cloud,” IEEE 3 rd international
VI. CONCLUSION conference on computer and Communication
We have implemented secured storage technology, 2012
system using symmetric key algorithms. Our 9. KhengKok Mar, “Secured virtual diffused file
experimental results shows that many network system for the cloud,” 6 th International IEEE
parameters like connection time, Handshake Time conference on internet technology and
and network load affects processing of files in secured transactions, UAE, December 2011
cloud environment. Proposed system deals with 10. RajkumarBuyya, Introduction to the IEEE
Transactions on Cloud Computing, IEEE
splitting of file into multiple chunks and storing
Transactions on Cloud Computing, Vol. No. 1,
parts in multiple cloudswhich solves the file January –June 2013
protection from malicious user but again the 11. NirnayGhosh, SoumyaGhosh, Sajal Das,
major problem with the system resources “SelCSP: A framework to facilitate selection of
requirement is too high and dependence of cloud service providers,” IEEE Transactions on
different cloud storage service provider come Cloud Computing, Vol. 3, No. 1, January-March
across. In near future work we proposed a system 2015
to deploy malicious user attack protection over 12. Chien-An Chen, Myounggyu Won, RaduStoleru,
GeofferyXie, “Energy-Efficient fault-tolerant
𝗂382 IJRAR- International Journal of Research and Analytical Reviews Research Paper
[VOLUME 6 I ISSUE 1 I JAN. – MARCH 2019] e ISSN 2348 –1269, Print ISSN 2349-5138
http://ijrar.com/ Cosmos Impact Factor 4.236
data storage and processing in mobile cloud,” 15. Sancha Pereira, Andre Alves, Nuno Santos,
IEEE Transactions on Cloud Computing, Vol. 3, Ricardo Chaves , "Storekeeper: A Security-
No. 1, January 2014 Enhanced Cloud Storage Aggregation Service",
13. Quanlu Zhang, Shenglong Li, Zhenhua Li, IEEE 35th Symposium on Reliable Distributed
Yuanjian Xing, Zhi Yang, Yafei Dai, “ CHARM: A Systems, 2016
Cost-efficient multi cloud data hosting scheme 16. Hussam Abu-Libdeh, Lonnie Princehouse,
with high availability,” IEEE Transactions on Hakim Weatherspoon, " RACS: A Case for
Cloud Computing, Vol. 3, Issue 3, July- Cloud Storage Diversity", International
September 2015 conference for Internet technology and
14. AlyssonBessani Miguel Correia Bruno Secured Transaction, December 2012
Quaresma Fernando Andre Paulo Sousa, " 17. Kevin D. Bowers, Ari Juels, AlinaOprea, "HAIL:
DEPSKY: Dependable and Secure Storage in a A High-Availability and Integrity Layer for
Cloud-of-Clouds", ACM Transaction on Cloud Storage", 16th ACM conference on
Storage, Vol. 9,No. 4, Article 12. November Computer and communications security,
2013 November 2009.

Research Paper IJRAR- International Journal of Research and Analytical Reviews 𝗂383