International Journal of Research in Advent Technology, Vol.7, No.

2, February 2019
E-ISSN: 2321-9637
Available online at www.ijrat.org

Secure Cloud Data Storage: From Single to Multi-Cloud

Environment
Manoj V. Bramhe, Dr. Milind V. Sarode, Dr. Meenakshi S. Arya
Phd Research Scholar, Professor and Head, Associate Prof. & Head
Department of Computer Science & Engineering, Department of Computer Engineering ,
G.H. Raisoni College of Engineering, Nagpur
Government Polytechnic, Yeotmal
manoj_bramhe@yahoo.com

Abstract: Cloud based storage services are most popular among organizations due to its high computing power at low cost with
saving in IT infrastructure. Major threat found by organizations before storing data in public cloud storage is security of the
information stored. Even after strong provision of various security mechanism at various cloud service level, still full security is
not achieved yet. Malicious system administrator can access data of any organization . Our proposed solution deals with the
cloud storage security issues by distributing trust and security of user's data among multiple clouds by distributing fragmented
data among them. Since no entity will get complete set of data at any instance of time hence system is secure and reliable.

Keywords: Cloud Computing, DFS, API

system, erasure coding, secret sharing, VM isolation and
1. INTRODUCTION cryptographic techniques for data storage. Main issues with
Cloud Computing saves operational and infrastructure cost such approaches is that data is available fully with public
for organizations as it provides networking, hardware and cloud provider where we cannot have trust on inside
software resources as per their usage but major problem system user as they can utilize data for malicious purpose
with public cloud storage systems is various security breaking security measures at IaaS level. Secondly some
challenges because whole user's data is available to service solutions were proposed for multi-cloud based system
provider. Various methodologies were proposed by many where security and trust is divided into multiple cloud
researchers for cloud security and privacy like hence dependency is less as compared to single cloud
cryptographic techniques, replication of data, trusted cloud systems. Authors in [6],[8],[9] has discussed approaches
computing, isolation of virtual machines etc but most of where data is divided into various parts and partial data
them were implemented in single cloud based system . parts are stored in various clouds. Various architecture for
Single cloud storages stores data at one place hence it faces data and application processing for multi-cloud
many problems like failure of service availability, environment is discussed in [4]. Authors in [7] have used
malicious system user, data integrity and data intrusion open source distributed file system named Tahoe-LAFS for
problems. Further research was carried into multi-cloud or reliability purpose. They have used secret sharing scheme
cloud-of-clouds where user's data is fragmented into and private key encryption methodologies. In [13]
multiple slices and distributed among various clouds. This researchers have proposed cost effective schema for
improves overall security, trust and reliability of storage storage in multi-clouds. Bessani et.al. have proposed
services as none of the entities get complete information at dependable storage clouds in [14] where they stores the
a time. Distributed file system (DFS) are used for users data partitioned in three clouds. They have used
managing files from multiple hosts. RAID techniques for recovery of the data. All of these
We have implemented multi cloud based storage approaches have focused on some specific security
service which not only provides security and privacy to parameters but our proposed system works on all security
user's data but also provide reliability. User's files are parameters.
divided into multiple chunks, encrypted and then stored in
multiple public/ private cloud in secured environment. 3. PROPOSED SYSTEM
Data is retrieved with integrity check and merged to We have studied solutions proposed by various researchers
generate original user file. Failure management is for single and multi cloud environment and concluded that
implemented through replication of data in multiple clouds multi-cloud systems are better options for storage systems.
. Multi-Clouds systems are more secure as they not only
2. RELATED WORK removes threats for single cloud systems like vendor-lock-
Cloud computing security research was carried in but also makes the data safe from malicious system user
out in two phases. In initial days solutions were proposed . Malicious activity is not possible for storage system user
for single cloud based system where as mentioned in [1] as he never have full copy of data available with him and
and [2]. They have used security approaches like trusted he cannot do anything with partial data.

625
 International Journal of Research in Advent Technology, Vol.7, No.2, February 2019
E-ISSN: 2321-9637
Available online at www.ijrat.org
Our proposed system takes dynamic decision as per end- 4. IMPLEMENTATION
user request, divide data followed by encryption to ensure We have developed system API for three main components
security. This encrypted blocks of data are stored in of the system for making it reusable in various
various cloud storage. The proposed system uses RAID applications. Independent dynamic linking libraries were
similar techniques to regenerate original file from partial developed for various methods as per the design of our
chunks in the event of failure of some cloud storages. DFS structure.
Figure 1 describes the architecture of proposed KeyGenaration, EncMethod, Encrypt, Decrypt are the
system. System is designed for multi-cloud environment functions written for encryption modules which are used
where application layer is developed as end user interface for processes required for maintaining security of files.
used by user for data and query provision. User requests SetFileCount and SetMode function are developed for De-
are processed by API developed for our system. We have construction module used for deciding no of fragments
used distributed file system mechanism for cloud storage . and mode of fragmentation. File Transport module uses
DFS command are useful for writing and reading data in functions developed like SetTransferMode, Connect,
cloud based system using its distributed characteristics. Disconnect, SendFileChunk, ReciveFilechunk for doing
DFS methods are developed for processing the files using network operation related to file writing and reading.
three main components of our systems as Encryption, De- These API sets allow end user to connect, configure and
Construction and File transport. utilize the storage system using FTP.
1) Encryption Component: It is designed for using
cryptographic techniques to ensure security of data before 5. EXPERIMENTAL RESULTS
storing in clouds. Encryption is used for partitioned file We have tested our implementation for single and multi
chunks before writing them into clouds and decryption is cloud based systems. We have used system configuration
used for getting original file along with merging operation as Intel P5 2,8GHZ processor with 2 GB RAM as local
during reading the files from clouds. We have used server . We have used different types and different size of
private key algorithms like AES and BlowFish for testing files for performance evaluation. System is tested for three
purpose. private key algorithms AES, DES and 3DES for
verification of their performances. We have considered
excel, doc, jpeg, html and pdf files of various size and
following graphs shows uploading time required for
various files in single and multi-cloud environment. Figure
2 shows time required for uploading various files using
AES encryption algorithm for single and multi-cloud
environment. Figure 3 and figure 4 shows uploading time
required for uploading various files using DES and 3DES
algorithm.
Our experimental results shows that DES
algorithm is slowest in performance and 3DES performs
better than DES whereas AES algorithm is best in class .
We conclude that AES is most suitable and fast private key
algorithm in cloud environment for encryption purpose
Figure 1: System Architecture

2)De-Construction Component: It is used for

splitting the original file into multiple parts and storing
them among various public cloud storages. System divides
whole file and stores first 1/3rd part as chunk 1 in local
server. this is used to store number of cloud storage server
required for the system. Further system divides remaining
parts in two chunks and stored them as chunk 2, chunk 3 in
two different public cloud storage.
3)File Transfer Component: It is used for writing
and reading file chunks over cloud environment using
TCP/IP framework. It uses socket programming and
network protocols like FTP. We have choose to use FTP Figure 2: File processing using AES
over HTTP as it is easy, fast and simple to implement. File
Transport Protocol uses port no 20 and 21 to write and
reads file to cloud storage server. Standard FTP client and
server management systems can be used for
implementation .

626
 International Journal of Research in Advent Technology, Vol.7, No.2, February 2019
E-ISSN: 2321-9637
Available online at www.ijrat.org
[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, Vol PP, Issue 99,
2013
Figure 3: File processing using DES [5]Tran Doan Thanh, Subaji Mohan, EunmiChoil,
SangBum Kim, Pilsung Kim “A Taxonomy and
Survey on Distributed File Systems,” IEEE Fourth
International Conference on Networked Computing
and Advanced Information Management, 2008
[6] Su Chen, Yi Chen, Hai Jiang, Laurence T Yang, Kuan-
Ching Li, “ A secure distributed file system based on
revised Blakely’s secret sharing scheme,” 11th IEEE
international conference on trust, security and
privacy in computing and communications, 2012
[7] Fan-Hsun Tseng, Chi-Yuan Chen, Li-Der Chou, Han-
Chieh Chao, “Implement a reliable and secure cloud
distributed file system,” IEEE international
symposium on intelligent signal processing and
Figure 3: File processing using 3DES communication systems, November 2012
[8] ShushantShrivastava, Vikas Gupta, Rajesh Yadav,
Comparing the results of various algorithms in Krishna Kant, “Enhanced Distributed storage on the
single and multi-cloud environment we conclude that small cloud,” IEEE 3rd international conference on computer
size files have similar uploading time for single and multi- and Communication technology, 2012
cloud environment whereas as the file size increases multi- [9] KhengKok Mar, “Secured virtual diffused file system
cloud based system gives better performance over single for the cloud,” 6th International IEEE conference on
cloud based system and required less time to upload data in internet technology and secured transactions, UAE,
storage server. Thus Multi-cloud based system are better December 2011
than single cloud based system. [10]RajkumarBuyya, Introduction to the IEEE
Transactions on Cloud Computing, IEEE Transactions
6. CONCLUSION on Cloud Computing, Vol. No. 1, January –June 2013
We have implemented system for secure data [11] NirnayGhosh, SoumyaGhosh, Sajal Das, “SelCSP: A
storage and evaluated its performance for single and multi- framework to facilitate selection of cloud service
cloud environment. Our system enhances security features providers,” IEEE Transactions on Cloud Computing,
in multi-cloud environment by distributing multiple file Vol. 3, No. 1, January-March 2015
chunks in various public clouds thus adversary never have [12] Chien-An Chen, Myounggyu Won, RaduStoleru,
complete information about user's data. We have evaluated GeofferyXie, “Energy-Efficient fault-tolerant data
working of system using AES, DES and 3DES storage and processing in mobile cloud,” IEEE
cryptographic algorithms. Our experimental results shows Transactions on Cloud Computing, Vol. 3, No. 1,
that AES is most suitable and fast security algorithm for January 2014
cryptographic operations. We also conclude that multi- [13] Quanlu Zhang, Shenglong Li, Zhenhua Li, Yuanjian
cloud based system gives better performance than single Xing, Zhi Yang, Yafei Dai, “ CHARM: A Cost-
cloud based system for file processing. efficient multi cloud data hosting scheme with high
availability,” IEEE Transactions on Cloud Computing,
REFERENCES Vol. 3, Issue 3, July-September 2015
[1] ]Zhifeng Xiao and Yang Xiao, “Security and Privacy in [14] Alysson Bessani Miguel Correia Bruno Quaresma
Cloud Computing”, IEEE Communications Surveys & Fernando Andre Paulo Sousa, " DEPSKY: Dependable
Tutorials, March 2012 and Secure Storage in a Cloud-of-Clouds", ACM
[2]MohammedA. AlZain, Eric Pardede, Ben Soh, James A. Transaction on Storage, Vol. 9,No. 4, Article 12.
Thom, “Cloud Computing Security: From Single to November 2013
Multi-Clouds”, IEEE 45th Hawaii International [15] Sancha Pereira, Andre Alves, Nuno Santos, Ricardo
Conference on System Sciences, 2012 Chaves , "Storekeeper: A Security-Enhanced Cloud

627
 International Journal of Research in Advent Technology, Vol.7, No.2, February 2019
E-ISSN: 2321-9637
Available online at www.ijrat.org
Storage Aggregation Service", IEEE 35th Symposium
on Reliable Distributed Systems, 2016
[16] Hussam Abu-Libdeh, Lonnie Princehouse, Hakim
Weatherspoon, " RACS: A Case for Cloud Storage
Diversity", International conference for Internet
technology and Secured Transaction, December 2012
[17] Kevin D. Bowers, Ari Juels, Alina Oprea, "HAIL: A
High-Availability and Integrity Layer for Cloud
Storage", 16th ACM conference on Computer and
communications security, November 2009.

628