International Journal of Machine Learning and Computing, Vol. 8, No.

4, August 2018

Ensuring Data Security in Cloud Storage

Shubham Singh and Akhilendra Pratap Singh


Abstract—Cloud computing is one of the most prominent II. RELATED WORK
storage and computing platform being researched nowadays. It
Confidentiality, Integrity and Availability are the key
is amongst the most popular networks in the world and is seen as
next generation of computing technology. It provides various attributes for any data and these problems cannot be solved by
services to its users. Among them, one of the salient service a single security method. Liu discusses about the traditional
offered is cloud storage. But the major concern while using this technologies and capabilities of cloud and the newer
service is the integrity and privacy of stored data. Users require technologies which must be used for better security and
their data to be safe and private from any modification or privacy of data [5]. The importance of security of data stored
unauthorized access. Thus security becomes an indispensable
part of the data stored on cloud. This paper proposes a way to
in the cloud has been emphasized many a times, along with
maintain data privacy and integrity and guarantees that no one the measures needed to secure the data in the cloud. In [6],
except the authorized user can access the data. AES algorithm is used to provide security to the end users of
Cloud. An encryption key is used to encrypt the user files to
Index Terms—Cloud computing, integrity, privacy, security. secure the content of the files. The encrypted files can be
transferred over the network without worrying about the files
getting accessed by an unauthorized user. Al-Jaberi et al. main
I. INTRODUCTION focus is integrity verification along with privacy preservation
Cloud computing is defined as a network based computing using algorithms and protocols for cloud stored data where
which provides shared processing data & resources to its user Amazon S3 was used as the cloud storage provider [7]. RSA
when required. As per NIST, it is a model which enables partial homomorphic & MD5 algorithm are used in the
pervasive, on-demand access to shared resources that can be proposed model [8]. RSA algorithm is used for securing data
rapidly provisioned & released with minimal management and MD5 algorithm is used for integrity verification. The
effort [1]. It provides a user various capabilities for storing author enhanced the cloud security mechanism using AES
their data at the cloud server & processing it when required. 128 bit, 192 bit & 256 bit key encryption to secure the data in
The availability of cheap storage devices & computers, high the cloud, depending on the size of the files [9]. In [10], for
capacity networks, service oriented architecture etc. are securely transmitting & storing data in the cloud, encryption
behind the growth of this technology. & decryption, both are done at clients end using a single key.
Cloud computing offers various advantages like better Mahalle et al. used data encryption using hybrid encryption
hardware utilization, reduced cost, high scalability etc. [2] to algorithm to ensure data security such that even the
both the cloud service provider and its users, and is now a administrator of the cloud server does not have access to the
highly demanded service. But despite of such advantages, private data of the user [11]. The keys are generated on the
organizations prefer to adopt the methodologies that are basis of system time thereby making the overall system more
already successful and used by other organizations. This is secure. In [12], the author proposed a new security model with
because of the risks associated with cloud computing. double authentication mechanism to be implemented to
There are some privacy concerns in the services provided restrict unauthorized persons from getting the control of
by the cloud [3]. The cloud service provider can access the user’s data. The user can also select the encryption techniques
stored data any time and can modify or delete it. It can share he wants to use.
the stored information with third parties, if necessary, as People are afraid of sharing their data with an untrusted
permitted in their privacy policy. A solution to protect the cloud service provider. The proposed model uses encryption
confidentiality of stored data is the way in which user stores and obfuscation techniques to guarantee higher level of
data. User can encrypt data before storing it in the cloud to security and confidentiality of user’s data [13]. Prasad et al.
prevent unauthorized access. Also, in a cloud provider proposed the implementation of High Security Password to
platform, used by a large number of users, the data belonging ensure the security of user’s data [14]. Any modification in
to different users may reside on same data server which may data will be done only if the user provides the password sent
lead to information leakage when a user’s information is given to his mobile. This is more secure than the traditional security
to other [4]. There is also a problem regarding the ownership mechanism of using a single password for the whole profile
of the stored data i.e. if a user stores his data on cloud, who the and tasks related to that account. The author has proposed
owner of stored data is? secure sharing functionality to be added for cloud service
providers using cryptographic algorithms like AES and RSA,
Manuscript received May 20, 2018; revised July 8, 2018. by associating particular permission decided by owner with
The authors are with the National Institute of Technology Meghalaya, generated keys that can be used to access a resource [15]. The
Shillong, India (e-mail: [email protected], resource can only be accessed if the key used by the user has
[email protected]).

doi: 10.18178/ijmlc.2018.8.4.716 382

 International Journal of Machine Learning and Computing, Vol. 8, No. 4, August 2018

the required permissions associated with it. The proposed compromised.

model uses a combination of authentication technique (digital This paper put forwards a new approach which ensures the
signature) and key exchange algorithm (Diffie Hellman) privacy along with the integrity of the data stored at cloud
blended with an encryption algorithm (AES) referred to as server.
“Three-way mechanism” to be used to protect confidentiality
of data stored in the cloud [16]. In [17], different security
issues arising from the usage of cloud services are discussed. IV. PRELIMINARIES
The author has proposed if one wants to take the cloud A way to provide security to data is cryptography which
computing to the next level, the security capabilities need to uses logic and mathematics to serve the purpose. Advanced
be strengthened. Elliptic curve cryptography encryption Encryption Standard (AES) is such a cryptographic technique,
technique is used for securing the data present in the cloud so a symmetric block cipher which performs all of its
that only authenticated user can access the data [18]. Any computations on bytes [22]. The number of rounds involved
breach at the cloud would not affect the security of stored data are variable & depends on key length. AES is used by a
as the data is in encrypted form. The author has covered all the number of organizations as it is more secure than other
cloud security techniques & countermeasures along with the algorithms and is faster in both hardware and software.
attacks possible against cloud based services and the Encryption and decryption of data is easy and gives good
countermeasures to thwart these attacks [19]. Mohamed et al. performance. It supports large key sizes of 192 and 256 bits
implements a software on Amazon EC2 Micro instance to along with its 128 bit form for heavy encryption purposes.
enhance data security in [20]. Also, the author gives three AES is considered invulnerable to all security attacks except
suggestions for Amazon EC2 users: algorithms like blowfish brute force. When its most secure 256 bit key is used, it will
or DES can be used for better performance, AES must be used take a large number of years to guess the key by brute force
if better security of user’s data is needed, and AES suits attack and hence it is almost unbreakable [23].
Amazon EC2 by providing better performance and security. Encryption itself is not sufficient for ensuring data’s
integrity. When a user stores his data on cloud, he needs an
assurance that it must remain intact. Therefore, Secure Hash
III. PROBLEM STATEMENT Algorithm 2 (SHA-2) is used, which is designed to protect the
There are several security issues associated with cloud integrity of data by detecting any modifications and warns the
storage like loss of data, loss of control, invalid storage etc. owner about it. It is a set of cryptographic hash functions
[21]. They are a major obstruction in adopting cloud which run on user’s data by comparing the computed hash
computing services as the confidentiality and integrity of value to a previously known hash value. Based on the digest
stored data is very important. Once a user stores his data on length, there are six hash functions in SHA family. Their shift
cloud, there must be some assurance that the data can be amount is different but they have identical structure which
accessed only by the authorized user and will remain differs in number of rounds [24]. SHA-2 is a stronger hashing
confidential. But due to some internal errors or malicious algorithm in comparison to other and is more suitable for
changes, user’s data might be exposed to unauthorized verifying the integrity of data as it is less susceptible to most
individuals or given to authorized users with integrity being cryptanalytic attacks.

Fig. 1. Proposed model.

V. PROPOSED SCHEME securely over the network. For it, AES algorithm is used as it
Fig. 1 shows the architecture of proposed scheme. It has the computational efficiency and speed to encrypt large
supports client side encryption so that data is transmitted data. The proposed algorithms are as follows:

383
 International Journal of Machine Learning and Computing, Vol. 8, No. 4, August 2018

ALGORITHM #1: AES Encryption b. select_file();

F: Selected file c. upload_server();
F_NAME: File name d. goto 5;
EF: Encrypted file 6. else if ( file_download ) then
IVAL: Hash value of encrypted file a. choose_file();
N: Number of encrypted file parts b. download_file();
EF_PART_NUM: Name of a part of encrypted file c. goto 5;
K: Encryption key 7. else if (logout())then
1. start a. goto 8;
2. F=select_file(); b. end if
3. K=enter_key(); 8. end
4. EF=AES_encrypt(F,K);
5. IVAL=find_SHA_val(EF); ALGORITHM #4: AES Decryption
6. insert_db(F_NAME,IVAL); F: Selected encrypted file
7. if ( split file in parts) then MF: Merged file
a. file_split(EF); Fi: Temporary file
b. for NUM=1 to N do K: Decryption key
i. IVAL=find_SHA_val(EF_PART_NUM); DF: Decrypted file
ii. insert_db(EF_PART_NUM,IVAL); N: Number of encrypted file parts
iii. output EF_PART_NUM ; EF_PART_NUM: Name of a part of encrypted file
iv. NUM++; 1. start
v. end for 2. K=enter_key();
c. goto 9; 3. if (file divided in parts) then
8. else a. MF=file_merge();
a. output EF; b. Fi=MF;
b. goto 9; c. goto 5;
c. end if 4. else
9. end a. Fi=F;
b. goto 5;
ALGORITHM #2: File Division c. end if
F: Selected file 5. if (check_integrity(Fi)) then
FS: Selected file size a. DF=AES_decrypt(Fi,K);
PS: Size of a part of the file
b. output DF;
RL: Total number of bytes to be read from the selected file
c. goto 7;
READ: Stores the bytes read from the selected file
IVAL: Hash value of encrypted file part 6. else
F_PART_NUM: New file part a. if (file divided in parts) then
1. start i. output “Modified file parts”;
2. F=select_file(); ii. for NUM=1 to N do
3. PS=input_partSize(); 1. if(!check_integrity(EF_PART_NUM)) then
4. FS=F.length(); a. output EF_PART_NUM;
5. if FS>0 then b. NUM++;
a. goto 7; 2. else
6. else a. NUM++;
a. output “Blank file”; b. end if
b. goto 8; 3. end for
iii. goto 7;
c. end if b. else
7. while FS > 0 do i. output “File modified”;
a. if FS <= PS then ii. goto 7;
i. RL=FS; iii. end if
b. else c. end if
i. RL=PS; 7. end
ii. end if
ALGORITHM #5: File Merge
c. READ=byte_read(F,0,RL); F: File name
d. FS-=RL; L: Array list
e. F_PART_NUM=create_file(); N: Number of file parts
f. write_file(F_PART_NUM,READ); F_PART_NUM: A part of the file
g. IVAL=find_SHA_val(F_PART_NUM); F_M: Merged file
h. insert_db(F_PART_NUM,IVAL); READ: Stores the bytes read from the selected file
8. end 1. start
2. F= enter_fileName();
ALGORITHM #3: Cloud Server Process 3. L=create_newList();
1. start 4. for NUM=1 to N do
2. client_login(); a. L.add(F_PART_NUM);
3. if (!client_login_success()) then b. end for
a. output ”Incorrect credentials”; 5. F_M= create_file();
b. goto 2;
6. for each F_PART_NUM in L do
4. else
a. READ=read(F_PART_NUM,0,F_PART_NUM.length())
a. goto 5;
b. write(F_M,READ);
b. end if
c. end for
5. if (file_upload) then
a. choose_server(); 7. end

384
 International Journal of Machine Learning and Computing, Vol. 8, No. 4, August 2018

VI. RESULT AND PERFORMANCE ANALYSIS useful for data whose protection is required because of the
The proposed model is successfully implemented using source or the nature of individuals involved, or when it is
java crypto package, javax.crypto, jre v1.8.0 and PHP v5.4.38 stored in a storage medium which is prone to risk of getting
on windows machine with Intel core i3 processor TM-4005U lost or stolen, and when it is to be transmitted through a
running at 1.70 GHz and 4G memory. network using mechanism which is not much secure. In [16],
As confidentiality of users data is most essential, therefore users data need to pass through two servers which is a major
to ensure it various cryptographic algorithms are there which drawback as it can be retrieved by some attacker during
transmission. Proposed scheme rectifies this problem by
can be used to protect the data from unauthorized access.
encrypting data at clients end such that data will be transferred
Although security is the most important factor, there are some
over the network only after getting encrypted which enhances
other factors which must be taken into consideration while
its security. The model in [10] stores complete data as a single
choosing an algorithm for encrypting the data, most important
piece at the server. If the server security is breached, data is
among which is the time taken for data encryption and
vulnerable. Proposed scheme allows the user to divide the
decryption. A comparison between some of the cryptographic
data in parts and store it at different cloud servers of his
algorithms is provided in [25] on the basis of time taken to
choice. In this way the user have the knowledge about the
encrypt a file. A text file of 50 MB in size is chosen as a
location of his data and if the security of any server is
sample whose encryption time by different algorithms is as
breached, data is still secure. If such an incident occurs and
depicted in Fig. 2. From the graph it is clear that AES some modifications are to the file, an integrity verification
algorithm takes the minimum execution time for data using SHA-2 informs the user about it.
encryption and decryption when compared with other
cryptographic algorithms. Therefore AES algorithm is the
best option when execution time is taken into consideration. VIII. CONCLUSION AND FUTURE WORK
The proposed scheme solves the problem of data security in
a distributed storage system and file division takes the
security to next level. User has privilege to store his data at the
server of his choice so that he’ll be able to track the server
where a security breach occurs. It provides a base for
enhancing the security of confidential data and future
enhancements can be made to it. A combination of various
encryption algorithms can be used which will make it almost
impossible for an unauthorized person to access the
confidential data. The model can further be improved by
incorporating schemes like PDP, ORUTA etc. to solve issues
Fig. 2. Encryption algorithm vs. time. like bandwidth, cost etc.

REFERENCES
[1] The NIST Definition of Cloud Computing. [Online]. Available:
http://csrc.nist.gov/publications/nistpubs/800-145/SP800-145.pdf
[2] H. Qusay, “Demystifying cloud computing,” The Journal of Defense
Software Engineering (CrossTalk), 2011.
[3] Cloud Computing Privacy Concerns on Our Doorstep. [Online].
Available:
http://cacm.acm.org/magazines/2011/1/103200-cloud-computing-pri
vacy-concerns-on-our-doorstep/fulltext
[4] A. Chhibber and S. Batra, “Security analysis of cloud computing,”
International Journal of Advanced Research in Engineering and
Applied Sciences.
[5] W. T. Liu, “Research on cloud computing security problem and
Fig. 3. File size vs. encryption time. strategy,” in Proc. 2nd International Conference on Consumer
Electronics, Communications and Networks, 2012, pp. 1216-1219.
[6] B. Thiyagarajan and R. Kamalakannan, “Data integrity and security in
The performance of an algorithm on same data of different cloud environment using aes algorithm,” in Proc. International
sizes is also important. Fig. 3 describes the time taken by AES Conference on Information Communication and Embedded Systems,
algorithm to encrypt a data file of different sizes [25]. It is 2014, pp. 1-5.
[7] F. A.-J. Mohammed and A. Zainal, “Data integrity and privacy model
clear from the graph that as the data size increases, encryption
in cloud computing,” in Proc. International Symposium on Biometrics
time of AES algorithm increases and vice versa. Thus, the and Security Technologies, 2014, pp. 280-284.
time taken by AES algorithm to encrypt a file is directly [8] P. Ora and P. R. Pal, “Data security and integrity in cloud computing
proportional to the file size. It means that a file having large based On RSA partial homomorphic and MD5 cryptography,” in Proc.
International Conference on Computer, Communication and Control,
size will take more encryption time then the file having less 2015, pp. 1-6.
size. [9] G. Raj, R. C. Kesireddi, and S. Gupta, “Enhancement of Security
Mechanism for Confidential Data using AES-128, 192 and 256bit
Encryption in Cloud,” Next Generation Computing Technologies
(NGCT), 2015 1st International Conference, pp. 374-378.
VII. DISCUSSION [10] N. Surv, B. Wanve, R. Kamble, S. Patil, and J. Katti, “Framework for
client side AES encryption technique in cloud computing,” in Proc.
The proposed model enhances the security of data. It is

385
 International Journal of Machine Learning and Computing, Vol. 8, No. 4, August 2018

IEEE International Advance Computing Conference (IACC), 2015, pp. [22] W. Stalling, Cryptography and Network Security: Principles and
525-528. Practice, 5th ed., 2011.
[11] V. S. Mahalle and A. K Shahade, “Enhancing the data security in cloud [23] Cryptography: What are the advantages and disadvantages of AES
by implementing hybrid (RSA & AES) encryption algorithm,” in Proc. over Triple-DES? [Online]. Available:
2014 International Conference on Power, Automation and https://www.quora.com/Cryptography-What-are-the-advantages-and-
Communication, pp. 146-149. disadvantages-of-AES-over-Triple-DES
[12] R. Kaur and R. P. Singh, “Enhanced cloud computing security and [24] Description of SHA-1 and SHA-256. [Online]. Available:
integrity verification via novel encryption techniques,” in Proc. http://www.quadibloc.com/crypto/mi060501.htm
International Conference on Advances in Computing, [25] R. Masram, V. Shahare, J. Abraham, and R. Moona, “Analysis and
Communications and Informatics, 2014, pp. 1227-1233. comparison of symmetric key cryptographic algorithms based on
[13] L. Arockiam and S. Monikandan, “Efficient cloud storage various file features,” International Journal of Network Security & Its
confidentiality to ensure data security,” in Proc. International Applications, vol. 6, no. 4, July 2014.
Conference on Computer Communication and Informatics, 2014, pp.
1-5
[14] M. S. G. Prasad, H. R. Nagesh, and L. Dharmanna, “Ensuring data Shubham Singh received the B.Tech degree in
storage in cloud computing for distributed using high security computer science and engineering from Invertis
password,” in Proc. National Conference on Research & Technology University, Bareilly, India, in 2015 and the M.Tech
in the Coming Decades, 2013, pp. 1-4. degree in computer science and engineering from the
[15] K. Chachapara and S. Bhadlawala, “Secure sharing with cryptography National institute of Technology Meghalaya,
in cloud computing,” in Proc. 2013 Nirma University International Shillong, India, in 2017.
Conference on Engineering (NUiCONE), pp. 1-3. He joined the Department of Computer Science
[16] P. Rewagad and Y. Pawar, “Use of digital signature with diffie hellman and Engineering, National Institute of Technology
key exchange and AES encryption algorithm to enhance data security Meghalaya, Shillong, India as a junior research
in cloud computing,” in Proc. 2013 International Conference on fellow, in 2017. He has authored a paper in a
Communication Systems and Network Technologies, pp. 437-439. peer-reviewed conference. His current research interests include cloud
[17] M. Z. Meetei and A. Goel, “Security issues in cloud computing,” in computing and hybrid automata.
Proc. 2012 5th International Conference on BioMedical Engineering
and Informatics.
[18] A. Kumar, B. G. Lee, H. J. Lee, and A. Kumari, “Secure storage and Akhilendra Pratap Singh was born in India. He has
access of data in cloud computing,” in Proc. 2012 International completed his Ph.D in Service Oriented Architectures
Conference on ICT Convergence, pp. 336-339. for wireless sensor networks from Indian Institute of
[19] M. Hamdi, “Security of cloud computing, storage, and networking,” in Information Technology, Allahabad.
Proc. 2012 International Conference on Collaboration Technologies Currently Dr. Singh is serving in the Department of
and Systems, pp. 1-5. Computer Science and Engineering at National
[20] E. M. Mohamed, H. S. Abdelkader, and S. EI-Etriby, “Enhanced data Institute of Technology, Meghalaya as an assistant
security model for cloud computing,” in Proc. 2012 8th International professor. His research interests include service
Conference on Informatics and Systems (INFOS), pp. CC12-CC17. oriented architectures, wireless sensor networks and machine learning.
[21] J. Grover, Shikha, and M. Sharma, “Cloud computing and its security
issues – A review,” in Proc. 2014 International Conference on
Computing, Communication and Networking Technologies, pp. 1-5.

386