Implementation of Deduplication on Encrypted Big-data using Signcryption for cloud storage

14 Pages·2017·0.49 MB·English

International Journal of Pure and Applied Mathematics
Volume 119 No. 12 2018, 13409-13421
ISSN:1314-3395(on-lineversion)
url:http://www.ijpam.eu
SpecialIssue
ijpam.eu
Implementation of Deduplication on Encrypted Big-data using Signcryption
for cloud storage applications
Saravanan Palani*, Sangeetha E, Archana A
School of Computing, SASTRA Deemed University, India
*Corresponding Author

ABSTRACT applying this technique, identical plaintexts
would turn out the same ciphertext.
As Big Data Cloud storage servers are
getting widespread the shortage of disc Keywords: Big data, digital signature,
space within the cloud becomes a major cloud, data deduplication, convergent
concern. The elimination of duplicate or encryption
redundant data, particularly in computer data
1. INTRODUCTION
is named deduplication. Data deduplication
is a method to regulate the explosive growth
Cloud Storage is an online storage service
of information within the cloud storage,
which provides services such as data
most of the storage providers are finding
maintenance, data management, and data
more secure and efficient methods for their
backup. The user is permitted to store their
sensitive method.
file online and can access the stored files
from anywhere. A recent survey states that
Recently, a noteworthy technique referred to
some business organizations gained an
as signcryption has been proposed, in which
advantage due to cloud adoption is nearly
both the properties of signature (ownership)
double the previously shown records, and it
and encryption are simultaneously
is estimated that by 2017, the cloud services
implemented with better performance
business may exceed $244 billion.
According to deduplication, we introduce a
Some of the most commonly known cloud
method that can eliminate redundant
services are Dropbox, Google drive, etc.
encrypted data owned by different users.
Stored files may be accessed from anyplace
Furthermore, we generate a tag which will
via net association. Each cloud service
be the key component of big data
allows a specific bandwidth. If a corporation
management. We propose a technique called
exceeds the given bandwidth, extra charges
digital signature for ownership verification.
would be applied. Some suppliers permit
Convergent encryption also called for a
unlimited information storage. It is one
content hash key cryptosystem. Convergent
among the issue that corporations ought to
encryption is an encryption approach that
take into account when looking at a cloud
supports deduplication. With this encryption
storage provider [15-24].
technique, the encryption key is generated
out of a hash of plain text. Therefore
13409
International Journal of Pure and Applied Mathematics Special Issue
Privacy is a major concerned factor one has server eliminates duplicated ciphertexts
to look for. Cloud encryption is the which ultimately improves the privacy
conversion of a cloud user’s data into protection. In this model, the security of the
ciphertext [25-32]. The Cloud storage ciphertext is improved in a better way
providers provide services like cloud compared with the existing encryption
encryption which is an encryption of the models. Usually, the encryption keys are
user's data before stored in the cloud [33- hash values of plain text, but here the
39]. Encryption is the scrambling of user’s encryption keys are randomly chosen. Here
data into a form such that it is impossible to the cloud server has no information except
decrypt the data without the the hash value.
acknowledgment of the cryptography key.
DupLESS: Server aided encryption for
Data security can be effectively achieved by deduplicated Storage [3]. They have
encryption. Encryption and secure proposed an architecture system called
encryption key management allow only DupLESS in such a way that the proposed
authorized users to access the uploaded data. architecture can resist the brute force attacks
The encrypted data is meaningless without through secure deduplicated storage. The
its respective key. messages were encrypted via PRF protocol
using the key which was obtained from the
2. RELATED WORK
server.The duplication check was done by
an existing service on client behalf.The main
Hybrid data deduplication in cloud
advantage of these techniques is that brute
environment [1].They propose a mechanism
force attacks are avoided, and clients can
called hybrid data deduplication. The
encrypt their data with a key server which is
solution for the mechanism mentioned
different from the separate storage server.
above is provided with partial semantics
The main drawback of this technique is that
security. In this system, CSP knows the
flexibility to other data users cannot be
encryption key of the data. This system
provided
model is not applicable where the Cloud
storage supplier cannot be trusted. The main Policy-based de-duplication model was
advantage of this technique is that it proposed based [4] on trust relationship
supports deduplication on plaintext and enabled on cloud storage components, de-
ciphertext. The main disadvantage of this duplication related components, and
technique is that it does not support different security requirements. The
encrypted data deduplication. In this proposed system model has a key
scheme, they use client-end de-duplication, management mechanism based on proxy re-
and the encryption is done at the client. In encryption algorithm for decryption of
this way, the computation pressure can be deduplicated data and data access. Even if
shared by all the clients. the malicious party tries to access the data
store in the cloud, it can’t gain knowledge
Encrypted Data Deduplication in Cloud
about Map box and the lockbox decryption
Storage [2]. The proposed model had cloud
13410
International Journal of Pure and Applied Mathematics Special Issue
private key. In case the chunk key is known, security’s standard notion for data
it may result in information leak. confidentiality was achieved. Key
management is another major challenge in
Efficient Deduplication On encrypted Big
cloud storage. In [8], security matrix and
Data in Cloud [5]. They have presented an
magic cards are used to identify right binary
efficient scheme, where the data was
pattern from the input stream.
encrypted using proxy re-encryption
technique and possession undertaking, then A Hybrid Cloud Approach for Secure
stored in the cloud.The outcomes display Authorized Deduplication [9]. The
scheme effectiveness and good performance authorized data deduplication issue was first
in deduplication of cloud storage data. addressed in this paper. Other than the data,
user’s different privileges were considered
Deduplication based on Hadoop [6].In this
for duplication check. The proposed model
paper, an integrated deduplication approach
was a hybrid cloud architecture which
is proposed by taking the features of Hadoop
supported authorized duplication check [10].
into account and leveraging parallelism
The proposed scheme had minimal
based on Map Reduce and HBase to speed
overhead, reduced storage space and saved
up the deduplication procedure. In our
network bandwidth. Key management
proposed approach, a new small-file
overhead issue was not addressed in the
aggregation scheme is proposed, and the
paper.
new standard of Secure Hash Algorithm-3,
Keccak is employed. The overall A Secure Client Side Deduplication Scheme
deduplication procedure is implemented in Cloud Storage Environments [11].
based on Map Reduce and HBase to provide OpenStack Swift was proposed for secure
scalability to cater to the challenges brought storage and data sharing via the public
by the big data Era. cloud. It is client-side deduplication.
Unauthorized user confidentiality was better
When there are too many files to process,
ensured. That is the user itself computed the
the tar-to-seq tool's performance may
key required for the respective data
sharply downgrade due to performance
encryption [12]. The data owner managed
bottleneck arising from memory or I/O in a
the data access. An authorized user can
single node.
decrypt a file in its encrypted form with the
private key only after integrating access
Attribute-Based Storage Supporting Secure
rights. The proposed model was successful
Deduplication of Encrypted Data in Cloud
only for hash key encryption. Deduplication
[7]. An attribute-based storage scheme was
was also used in cloud storage auditing [13]
proposed in a hybrid cloud setting for secure
and improving the access mechanism with
deduplication. The public cloud is
higher efficiency [14].
responsible for storage. The private cloud
manages the duplication check. Confidential
The proposed model in the paper is done by
data sharing was done efficiently by
using convergent encryption and digital
specifying access policies. Semantic
signature verification for deduplication on
13411
International Journal of Pure and Applied Mathematics Special Issue
big encrypted data. Confidentiality of data 2) Cloud Storage Provider is used to store
was better ensured. The proposed model the encrypted data. In this scheme, CSP is
provided good performance in achieving only used for storage as it cannot be trusted
deduplication on encrypted big data. fully. The trust issue is due to the curiosity
of the cloud to know about the stored data.
3. PROPOSED APPROACH
3) The authorized party is used for data
Popular service of the cloud is storage of ownership verification and duplication
data. Respecting the privacy of the user, the checking. The data holders completely trust
data will be encrypted first and then stored AP.
in the cloud. New challenges arise for data
deduplication in the cloud due to the 3.1. PRELIMINARY AND NOTATION
encrypted data storage which leads to
3.1.1PRELIMINARIES
difficulties in big data storage.
1)Convergent Encryption
Here the convergent encryption technique is
represented as (KG; E; D)Where KG is the
key generation, E is the encryption, D is the
decryption. On the input of the hash key K
and the message M, the encryption
algorithm E generates a ciphertext C where
C=E (M, K). On the input of hash key K and

the ciphertext C, the decryption algorithm
Fig 1. Proposed model generates a plain text M where M=D(C, K)
2).Symmetric Encryption
In this paper, the scheme is proposed by The encryption algorithm takes two
applying CE based on digital signature parameters namely the plaintext M and the
verification. In the proposed scheme, it is secret key K. Using these parameters, it
not necessary for the data holders to be in encrypts the input M with the secret key S
online for the duplication check. and generates the ciphertext CT (I.e.) E (M,
K). This process is conducted at the
The scheme has three entities. They are
respective data owner U; then the encrypted
data is uploaded to the cloud storage
1) Data holder who would upload the data
provider CSP
which will be saved in the cloud. In this
scheme, it’s possible that the data holders
The decryption algorithm takes two
will try to upload the same data which is
parameters namely the ciphertext CT and the
already stored in the cloud storage in its
secret key K. Using these parameters, it
encrypted form by a data holder.
decrypts the ciphertext CT with the secret
13412
International Journal of Pure and Applied Mathematics Special Issue
key S and generates the plaintext M (I.e.) D 2) Data duplication- when a data holder
(CT, K). This process is conducted to obtain tries to store a data which is already stored
the encrypted data stored in the cloud in the in the cloud storage, duplication occurs. AP
form of plain text does the duplication check using tag
comparison. If the comparison results in
3.1.2 NOTATION
positive, AP contacts the data holder and
informs about the duplication occurred
Table 1 represents the list of parameters
used in the proposed model
3) Data download- If the data holder
wants to decrypt a particular data stored at
Key Description
M The plaintext CSP, it will request AP to provide the key
CT The ciphertext
needed to decrypt the respective data. AP
H() The hash function
KG The Key generation goes for data holder eligibility check. If the
algorithm check results in positive, then AP will issue
K The hash key
the secret key that will be used to decrypt
T() The tag generation
function the encrypted data by the respective data
T The tag used for data
holder
duplication check
S The secret key
(PK1, Pk2) The public key and 4.1DATA OWNERSHIP VERIFICATION
private key for SCHEME
ownership verification
E The encryption
In this paper, the data ownership verification
D The decryption
SIG The digital signature is done based on Rivest-Shamir-Adleman
HD The hashed document scheme. AP checks data holder’s eligibility
Encrypt(M,S) The encryption function
to make sure that it is a real party which will
on M using secret key S
Decrypt(CT.S) The decryption function upload the data. Here H (M) cannot be
on CT
disclosed by CSP.
CS(HD,PK2) The signature creation
function
VS(HD,SIG,PK1) The signature 4.2 PROCEDURE
verification function
4.2.1 Data deduplication
4. SCHEME The procedure for data deduplication
is given by the following steps:
The proposed model has the following
aspects Step 1) user U1 selects the data
M it wants to upload
1) Data upload- AP does the
duplication check. If the check results in Step 2) Tag generation for the
negative, then the data holder encrypts the particular data is done by user U1. It will
data with the secret key S given by AP. The generate a tag for the particular data M, t1=
encrypted data is then stored at CSP H (E (H (M), M))
13413
International Journal of Pure and Applied Mathematics Special Issue
Step 3) AP will do the
duplication check. It will check whether the
tag t1 already exist
If the check results in negative, then the
data holder will encrypt the data M with the
secret key S given by AP. The encrypted
data is then stored at CSP. If the check
results in positive, then AP contacts the data
holder and informs about the duplication
occurred

Step 4) When User U2 request
Fig 2. Data ownership verification
to upload the same data M encrypted by the
user U1 following step 1 and 2. In step 3,

duplication happens because t2 exists so that
U2 will be informed about the situation. Now, the digital signature is created for the
file using the hashed document and the
4.2.2 Ownership verification
private key. AP does the verification of the
signature using its respective public key. If
Step 1) the public key and the private key
the verification is true, the AP proceeds to
(PK1, PK2) is generated for each message
check for duplication
which has to be uploaded by the data holder.
Step 2) Signature creation: The signature is
created on the input of hashed message
(HD) using private key PK2
Step 3) AP verifies the signature on the
input of hashed message HD, signature SIG
using public key PK1
Step 4) if the verification is positive, AP
proceeds to check for duplication. If the

verification is negative AP will conclude it
as an attempt by a malicious party. Fig 3. Data deduplication
4.2.3 Data ownership verification 4.2.4 Data deduplication:
The data owner logs into its account. It will The data owner logs into its account and
select the document that it wants to upload selects the document that it wants to upload
to the cloud. Then, the document is hashed. to the cloud. A digital signature is created
Public and private key pair is generated for for the particular document, and AP verifies
the selected document. it. If the verification is true, then AP checks
13414
International Journal of Pure and Applied Mathematics Special Issue
for duplication. The process is done by 1)
deriving a hash key from the selected
document. 2) The ciphertext of the particular
document is done by encryption of the
selected document and the hash key. 3) A
tag is generated from the corresponding
ciphertext created. 4) The created tag is
compared with the other tags present in AP
if the comparison is false, then it is uploaded

to the cloud
Fig 5. Data Decryption

It will request AP to share the secret key for
the selected document by sending its digital
signature for verification. AP verifies the
digital signature. If the verification is true,
then AP shares the secret key to the
respective data holder. Now the data holder
can decrypt the selected document using the
Fig 4. Data Encryption
respective secret key.

5. PERFORMANCE EVALUATION
4.2.5 Data Encryption:
The proposed model was implemented and
The data owner logs into its account and tested for performance evaluation. The time
selects the document that it wants to upload taken for selection of data to be uploaded
to the cloud. A digital signature is created and downloaded were not taken into
for the particular document, and AP verifies consideration. Testing was mainly focused
it. If the verification is true, then AP checks on algorithms designed and deduplication
for duplication. If the comparison is false, procedure performance
the AP generates a secret key. The data
owner encrypts the selected document with 5.1 Data Encryption and Decryption:
the secret key and then uploads it to the
Data encryption and decryption were done
cloud.
by using AES. The operation time was
estimated using AES key size 256 bits for
4.2.6 Data decryption
different data size (range from 80MB to
A data holder logs into its account, it will 150MB). It was observed that the time taken
select the document that it wants to decrypt. for the data which was as big as 150MB to
encrypt and decrypt were 64 milliseconds
and 60 milliseconds respectively.
13415
International Journal of Pure and Applied Mathematics Special Issue
5.3 Data ownership verification:
Fig 6. Data Encryption and decryption

Fig 8. Data ownership verification
As the size of the data increased, the time is
For each step in data ownership verification,
taken for data encryption and decryption
the operation time was estimated using 2048
also increased. Data encryption and
bit RSA. For a data of 100MB, the time is
decryption were efficiently done by using
taken for key generation, signature creation
AES.
and signature verification were observed.
5.2 Convergent Encryption: The time taken for a key generation was
11milli second. The time taken for creation
of signature was 16 milliseconds. The time
taken for verification of signature was 15
milliseconds. In the proposed model, the
data holder didn’t need to proceed for data
encryption if that data was already uploaded
by another data holder
References
Fig 7. Convergent Encryption
[1]. C. Fan, S. Y. Huang, and. C. Hsu,
Hybrid data deduplication in a cloud
Each operation of convergent encryption
environment, in Proc. Int. Conf. Inf.
was tested under different AES key size
Secure. Intell. Control, 2012, pp.
128bits, 196bits, and 256bits. For a data of
174177
100MB, the time is taken for key generation,
encryption, tag generation, decryption were
[2]. Encrypted Data Deduplication in
observed. For the different AES key size,
Cloud Storage, Information Security
the encryption time was less than 44
(AsiaJCIS), 2015 10th Asia Joint
milliseconds. The decryption time was less
Conference on, Issue Date: 24-26 May
than 41 milliseconds. For each operation, the
2015
computation time did not have much
difference. The proposed model was [3]. M. Bellare, S. Keelveedhi, and T.
efficiently applicable in various scenarios. Ristenpart, “DupLESS: Server aided
13416
International Journal of Pure and Applied Mathematics Special Issue
encryption for deduplicated Storage,” Transactions On Parallel And
in Proc. 22nd USENIX Conf. Secur., Distributed Systems, Vol. 26, No. 5,
2013, pp. 179–194. May 2015
[4]. Chuanyi Liu, Xiaojian Liu and Lei [10]. C. Fan, S. Y. Huang, and W. C. Hsu,
Wan “Policy-based deduplication in “Hybrid data deduplication in cloud
secure cloud storage,” in Proc. environment,” in Proc. Int. Conf. Inf.
Trustworthy Compute. Serv., 2013, pp. Secur. Intell. Control, 2012, pp. 174–
250–262, doi: 10.1007/978-3-642- 177, doi:10.1109/ISIC.2012.6449734.
35795-4_32.
[11]. Nesrine Kaaniche, Maryline Laurent
[5]. Y.Saritha, A. Vineela, "Efficient “A Secure Client Side Deduplication
Deduplication On encrypted Big Data Scheme in Cloud Storage
in Cloud”, International Journal of Environments” IEEE Transactions on
emerging trends and technology in Mobility and Security (NTMS) in
computer science, vol.6, no.4, 2017. Cloud Computing, Issue
Date:April.2.2014
[6]. Zhang, D., Liao, C., Yan, W., Tao, R.,
& Zheng, W. (2017, August). Data [12]. C. Yang, J. Ren, and J. F. Ma,
Deduplication Based on Hadoop. In “Provable ownership of file in
Advanced Cloud and Big Data (CBD), deduplication cloud storage,” in Proc.
2017 Fifth International Conference on IEEE Global Commun. Conf,
(pp. 147-152). IEEE. 2013,pp.695700,doi:10.1109/
GLOCOM. 2013.6831153.
[7]. Hui Cui, Robert H. Deng, Yingjiu Li,
and Guowei Wu “Attribute Based [13]. J. W. Yuan and S. C. Yu, “Secure and
Storage supporting secure constant cost public cloud storage
deduplication of encrypted data in auditing with deduplication,” in Proc.
cloud” IEEE Transactions on Big IEEE Int. Conf. Communic. Netw.
Data, January 2017 Secur., 2013, pp. 145–153,
doi:10.1109/CNS.2013. 6682702.
[8]. Sekar, K. R., Saravanan, P., &
Sethuraman, J. (2006). Discovery of [14]. T. Y. Wu, J. S. Pan, and C. F. Lin,
right binary pattern in key “Improving accessing efficiency of
management using security matrix and cloud storage using de-duplication and
magic cards. ARPN Journal of feedback schemes,”IEEE Syst. J., vol.
Engineering and Applied Sciences, 8, no. 1, pp. 208–218, Mar. 2014,
11(15), pp. 9187-9194 doi:10.1109/ JSYST.2013.2256715.
[9]. Jin Li, Yan Kit Li, Xiaofeng Chen, [15] Indragandhi, V., Logesh, R.,
Patrick P.C. Lee, and Wenjing Lou “A Subramaniyaswamy, V., Vijayakumar,
Hybrid Cloud Approach for Secure V., Siarry, P., & Uden, L. (2018).
Authorized Deduplication” IEEE Multi-objective optimization and
13417
International Journal of Pure and Applied Mathematics Special Issue
energy management in renewable electroencephalography feedback.
based AC/DC microgrid. Computers & Biomedical Research, 28(13), 5646-
Electrical Engineering. 5650.
[16] Subramaniyaswamy, V., Manogaran, [21] Arunkumar, S., Subramaniyaswamy,
G., Logesh, R., Vijayakumar, V., V., Karthikeyan, B., Saravanan, P., &
Chilamkurti, N., Malathi, D., & Logesh, R. (2018). Meta-data based
Senthilselvan, N. (2018). An ontology- secret image sharing application for
driven personalized food different sized biomedical images.
recommendation in IoT-based Biomedical Research,29.
healthcare system. The Journal of
[22] Vairavasundaram, S., Varadharajan,
Supercomputing, 1-33.
V., Vairavasundaram, I., & Ravi, L.
[17] Arunkumar, S., Subramaniyaswamy, (2015). Data mining‐based tag
V., & Logesh, R. (2018). Hybrid recommendation system: an overview.
Transform based Adaptive Wiley Interdisciplinary Reviews: Data
Steganography Scheme using Support Mining and Knowledge Discovery,
Vector Machine for Cloud Storage. 5(3), 87-112.
Cluster Computing.
[23] Logesh, R., Subramaniyaswamy, V., &
[18] Indragandhi, V., Subramaniyaswamy, Vijayakumar, V. (2018). A
V., & Logesh, R. (2017). Resources, personalised travel recommender
configurations, and soft computing system utilising social network profile
techniques for power management and and accurate GPS data. Electronic
control of PV/wind hybrid system. Government, an International Journal,
Renewable and Sustainable Energy 14(1), 90-113.
Reviews, 69, 129-143.
[24] Vijayakumar, V., Subramaniyaswamy,
[19] Ravi, L., & Vairavasundaram, S. V., Logesh, R., & Sivapathi, A.
(2016). A collaborative location based (2018). Effective Knowledge Based
travel recommendation system through Recommeder System for Tailored
enhanced rating prediction for the Multiple Point of Interest
group of users. Computational Recommendation. International
intelligence and neuroscience, 2016, Journal of Web Portals.
Article ID: 1291358.
[25] Subramaniyaswamy, V., Logesh, R., &
[20] Logesh, R., Subramaniyaswamy, V., Indragandhi, V. (2018). Intelligent
Malathi, D., Senthilselvan, N., sports commentary recommendation
Sasikumar, A., & Saravanan, P. system for individual cricket players.
(2017). Dynamic particle swarm International Journal of Advanced
optimization for personalized Intelligence Paradigms, 10(1-2), 103-
recommender system based on 117.
13418