Proceedings of the SMART-2016, IEEE Conference ID: 39669
5th International Conference on System Modeling & Advancement in Research Trends , 25th _27'h November, 2016
College of Computing Sciences & Information Technology, Teerthanker Mahaveer University, Moradabad , India
An Efficient Architecture and Algorithm to
Prevent Data Leakage in Cloud Computing using
Multi-tier Security Approach
Anshu Kirar l , Arun Kumar Yadav 2 and Supriya Maheswari 3
1,2Dept of Computer Science & Engg" ITM University Gwalior, Madhya Pradesh, India
3Dept of Computer Science & Engg" Hindustan College of Science & Tech., Mathura, Uttar Pradesh, India
E-mail : lanshukirar07@gmail.com.2arun26977@gmail.com.3supmaheswari@gmail.com
Abstract-Cloud Computing is arising on-demand,
ubiquitous internet based technology, which provides variety
of services. Cloud computing offers various services like
software, hardware, data storage, platform, infrastructure
etc. major weakness with data security over cloud is that, the
but of physical security control is not viable. Appropriately,
strong approach control and authentication becomes
extremely vital for providing valuable security. Number of
protection and privacy issues rise be-coming to the
outsourcing of infrastructure, delicate records and critical
application and its multitenancy environment. In this
research paper, we are proposing authentication technique
and multifactor user authentication for prevent illegal access
over cloud. As our proposed technique using three tier of
authen-tication check, authentication image check and
authentication codes. To provide further security on client
side, we are using a special keyboard and on the server side
authentication code is store in encrypted form on master
server at different tiers. Our proposed technique is quite
more secure.
Keywords: Data Leakage, Security, Cloud computing,
Secure Keyword
L INTRODUCTION
In present day the evolution and advancement in the
scientific knowledge make the world strong but it also
some dark side, There are several of the peoples who are
unsocial with the help of tool and they create harmful
special effects on our life, In this field there are various
security problems as it incorporates many technologies
together with memory mana-gement, load balancing,
resource scheduling, databases, net-works, virtualization
operating systems, transaction man-agement and
concurrency control.
These technology required security challenges at all
step on cloud, For example the network on which cloud is
running necessity is securing furthermore the virtual
server which handles the whole system must be safe and
require several securItIes, Communication between
physical machine and virtual machines must be securely
managed, Various policies to be maintain for data safety
Copyright © SMART-2016 ISBN: 978-1-5090-3543-4
and data encryption on the various steps, Malware
detection should be finding using data mining techniques,
If there is information store over cloud anyone can
access data from anyplace and anytime but these sensitive
data must be saving on client side along with on server
side, Therefore it is fundamental of certain data security
further data integrity process, Secondly, data thieving is a
one of sever dispute so it must be control on every single
step.
It is internet base computing that delivers pool
processing resources like server, storage, application,
services and net-works. Cloud computing control and
manage the resources to be focused on cross requirement.
It is continuous growing and performance evolutions
growth. But the major challenging task is the security and
privacy issues caused by outsourcing of infrastructure.
• Maintenance is simple and easier applications so
it is not required to be installing on every
computer so we can easily access from diverse
places.
• Reliability improves by the usage of multiple
spare sites, It provide offer a SLA service level
agreement which guarantee 365 and 99,99%
availability,
Security is well-set of control based technology and it
protects the infrastructure of cloud atmosphere, Cloud san-
ctuary is an enlarge sub sphere of computer security,
network security and largely information security,
A Security Risk in Cloud Computing
Top nine security risks are: L Data Breaches 2, Data
Loss 3. Account Hijacking 4, Insecure APIs 5, Denial of
Service 6. Malicious Insiders 7. Abuse of Cloud Facilities
8, Insufficient Due Diligence 9. Shared Technology Issues
Security Risk can be categorized into two:
• Risks for cloud service user
• Risks for cloud service provider
Data Leakage is distinct as unintentional distribution
of private or susceptible data to unauthorized person. Data
271
5th International Conference on System Modeling & Advancement in Research Trends, 25th _27'h November, 2016
College of Computing Sciences & Information Technology, Teerthanker Mahaveer University, Moradabad , India
Leakage event can be categorized into two kinds directs
losses and indirect losses. Direct losses refer to the
sustainable da-mage i.e. easy to evaluate to estimate
quantitatively. Indirect losses it is a big transaction to
enumerate as compare to direct loss and it also impact in
terms or cost place and time. According to survey of a
reputed company with the advance-ment of info
technology, there are accesses in the chance of cyber-
attacks. Cyber-attacks are the attacks in which a person
using an information of another person. Information like
his online bank details, Transaction details, Transaction
pass-words etc. and utilize them for his advantage without
notify the account controller. It come in to category of
cyber Attacks. Appropriate to this the account owner face
a very big loss whether the loss related to their money, his
personal details etc. To overcome this type of disaster we
work on a method in which multiple security checks is
happened. Due to this it minimizes the chances of data
lost. For maintaining the good secure access on online
transactions. We use Secured Key-board process. It offer
the best technique for this requirement based on Multiple
Security Checks, it can be execute on secure online
transactions.
That software packages also work with SET, SSL and
PKI, Secure E-commerce protocol technology for
encryption of data spreads. Online transactions operate
over on internet or intranet.
Secure Electronic Transaction (SET) is a
communications protocol. It is use for securing credit card
transactions over in tern et insecure networks. Security
protocols and their for-mats that are enable user to draw
on the available debit card premium organization on an
open network in a secure mode. SET key functions are
Confidentiality of information, In-tegrity of data,
Cardholder account authentication, and Merchant
authentication.
Secure Sockets Layer (SSL) is a typical security
technology. SSL establish an encrypted linkage among a
server and a client. It allow susceptible information like as
credit card, visa, social security information, and login
approval to be transmit securely. Data sent among search
engine and web servers in a plain text- quit you
susceptible to eavesdrop. If an attacker is capable to
interrupt whole data being sent among search engine and a
web server so it can be observe and use that information.
A public key infrastructure (PKI) is a bent of
characters and it manages public-key encryption. The
roles of a PKI are to make easy the secure electronic
transaction of information for an assortment of network
behavior like electronic commerce, in tern et banking and
confidential email.ltis crucial for activities, where
straight forward passwords is insufficient authentication
way and further accurate verification is ne-cessary to
verify the uniqueness of the parties involved in the
272
communication and to validate the information being
transferred.
Security is a necessary part of at all transaction that in
excess of the internet. Customer will lose his/her
confidence in e-business if its security is compromised.
The essential requi-rements for secure e-payments/
dealings confidential, integrity, availability, authenticity,
non Repudiablity, encryption, audit-ability.
11. REL ATED WORK
In today' s world the improvement and advancement
in the technology make the world strong according to the
technology but there is also there is some dark side. [1]
There are numerous of the peoples who are unsocial with
the help of technology who make us stronger. Now come
to the OTP technique it is simple but complicated but day
by day achievements in the technology of smart phone or
information transfer. We have face the OTP procedure to
protect the our delicate information from the impartial
peoples, who exploitation the information for their
benefits. An example if we had to do some transactions
with our bank whether, we use M-Banking, E-Banking
and when we initiate the preferred transaction, we get an
OTP on our RMN via SMS and return the OTP within a
squat interval to complete the cycle. To stay away from
any possible attacks like phishing, man-in-the middle
attack, mal ware Trojans, the OTP necessity to be secured.
In sort to make available a consistent and safe mode of
online transactions without any compromise to conv-
enience, a reliable m-banking authentication scheme that
combines the top secret PIN with encryption of the OTP
has been refmed in this paper. The combination of PIN
with OTP provides authentication and security.
The proposed method provide security even if any
dispute arise due any probable attacks like intern et
hacking or mobile thefts. The main impartial of this paper
is securing OTP using Festal Network Process. By using
this method we can easily change the size of the input we
provide. The sub-Keys are generated in each round
appropriate to this cascading iteration is produced. If we
use more round of encryptions it is difficult to break the
OTP.
A diverse approach has offered that raises the stage of
security dimensions using cryptographic techniques E-
com-merce portals, because day by day e-commerce had
created an enormous effect on the peoples. Nearly every
one of the peoples buy their goods from E-commerce
portal by online transaction process and due to this it is
very dangerous to maintain the transaction information of
the people would be in secure hands. For maintaining the
good secure access on E-commerce transaction [2] we use
(PKI) it offer the best approach for this requirement.
Based on PKI numerous security maintenance can be
implementing for secure E-com-merce transaction. E-
Copyright © SMART-2016 ISBN: 978-1-5090-3543-4
 An Efficient Architecture and Algorithm to Prevent Data Leakage in Cloud Computing using Multi-tier Security Approach
commerce software packages work with SET, Secure
Socket Layer, and Public Key Infrastructure and Secure E-
commerce protocol technology for encryption of data
transmission. E-commerce operates on in tern et or intranet.
Now existence we do lots of transactions like as Online
Recharge, Fund Transfer, bank transactions and m-
commerce etc. using smart phones and PDAs which
comes handy. Due to this all of the working made easy.
Information security has turn into a very critical aspect of
modern communication. When people perform a
transaction over internet, the security of information
against security threats is a major issue. It is an extremely
big issues that how we secure the information of the
community they provide. Appealing good privacy can be
used to offer authentication and confide-entiality to E-
commerce security.
In this research paper researchers deliberate the
problem [3] face by in tern et banking since that is tricky
issues of faith and safety of online Transactions. And we
deliberate that a number of superiority of client/public are
always afraid concerning the protection of their online
transactions and they are not capable to easily trust the
web fears that their online transactions may not be safe
payable to the increasing numbers of online intern et
attacks. To overcome this type of in tern et intrudes. Now
days a latest model for in tern et banking transactions is
represented. By using this model of transactions it
increases the security and trust above the accessible
model , by allowing customers and banks have to
communicate with one another and sign process
transactions online. The core opinion behind the conquer
the security issues that it enhance security by use of three-
tier, trusted, layered, and secure channel. This model
ensures that only liable people have the access to intern et
bank accounts suitable to this the information is remain
private and unable to modified by third parties and that the
transactions made are appreciable and verified.
Now these days all we techies familiar with a term of
Cloud Computing and it is based on internet computing
;wherever a fIrm of assets and services such as
applications, storage and servers are deliver to computers
and devices during the inter-net [4]. It incorporates large
open distributed system, virt-ualization, and intern et
release of services, dynamic provision of reconfigurable
resources and on demand operations. It is continuous
growing and performance regular evolution in the range of
computing. But the major challenging task is the safety
and confidentiality issues caused by outsourcing of
infrastructure. Delegate information and critical
applications and its multi-tenancy environment the volatile
growth of cloud computing has made the condition of
sufficient and valuable security challenges. Multi- factor
user authentication is an effective technique for preventing
un-authorized access. The main flaw in the security of
cloud is that, the arrangement of physical security controls
Copyright © SMART-2016 ISBN: 978-1-5090-3543-4
is impossible. As a result, strong access control and
authentication become very vital for provide effective
security. However, there are still other security issues to
be addressed in the future.
Today in the present world mostly depends on the
exchange of information i.e. the transmit of information
among more than two person; we call it as distributor
system . The data is sent to the highly confidential person
so the data allocation is just happened among the
distributer and the third party. The data sent by the source
would be secured and confidential and should not be
reproducing because the data are confidential and highly
important. At various point the information is dispatch or
copied by unsocial folks, who are liable for the data loss
and other type of damages to the system . These type
things come under process name data leakage. It must be
detected near the beginning of the stage in directive to
assure the data from being come in the presence of
unsocial individuals. So In this scheme we deal through
the data leakage by using several encryptions to the
susceptible data so it cannot be reproduced. The research
compacts with the thought of creating a dot arrangement
on a file at convinced site and while the arrangement is
completed it generates a wa-termark [5]. According to this
process data records are secure and protected with
unsocial folks.
In this paper author deliberate regarding to the data
security of the fortuitous data leakage [6]. It could be
happened when tenant organize applications beneath the
control of third-party service providers, they should trust
the provider' s security mechanism for tenant isolation,
resource sharing and access control. But in some of the
cases the unintentional data leakage may develop expected
to misconfiguration or virus in the cloud platform. To
overcome the issues a term Cloud Safety Net (CSN), an in
substantial monitoring framework that give tenant
visibility into the spread of their request data in a cloud
environment with low performance overhead. It exploits
the incentive of tenants to cooperate with each and every
to identify unintended data leakage. The central
persistence of CSN that it is an Tag based approach to
monitoring data flows is related to information flow
manage Techniques, encryption and digital water-
marking. With the assist of this we put a close sense on
unexpected data leakage and save the data to gone in the
wrong hands. Cloud security concerns in which they talk
about the procedure followed by the Austrian government
in which the information of the peoples who live there will
cross checked and upload on the cloud server. The process
is very simple if you are an authentic citizen of the country
then u have to assign online for the verification process
and when you clear that process you are registered and
your e ID (Electronic Identification) is generated.
The main benefit of this e ID system in several
situations, a centralized operation path of MOA-ID may
273
5th International Conference on System Modeling & Advancement in Research Trends, 25th _27'h November, 2016
College of Computing Sciences & Information Technology, Teerthanker Mahaveer University, Moradabad , India
be desirable. Therefore we propose a centralized
distribution approach of MOA-ID in the public cloud.
With the help of this categorized system data leakage and
loss is very squat and it always helps and creates trust
while doing online transactions through cloud services. [7]
This paper basically a challenge to learn secret data with
privacy leak detection in the context of text mining. Here
exists a facility to group these documents base on the
concept base mining algorithm. The group of documents is
always present in the form of hierarchy. When an operator
in the position of a subscriber desires to access a
document, this request for access will have to go through
an authentication procedure based on the Leakage Free
Redact able Signature Structure. All of this information
present in the form of cloud user access control list. A
privacy detection leak module it detects privacy leaks on
ahead the pattern of preceding privacy leaks is also being
protected. Appropriate to the above information is use to
update the cloud subscriber access manage list and those
are liable for data leakage is illegal to access the facility. It
is protest to organize the privacy protection with pattern
recognition approaches to data drip re-cognition in the
context text mining. In the future by integrating pattern
recognition as a feature of privacy presser-vation the cloud
service providers will be capable to preserve the privacy
of the susceptible data being shared.
In this paper [8] author discussed about list of
parameters that influence the security of the cloud and it
analyze the cloud security disputes and complication faced
by service provider and also cloud service consumer such
as data, privacy, and affect application and security
disputes. It also discusses some guidelines for accept these
issue and troubles. The service provider and the customer
should make certain that the cloud is safe enough from all
the external threats, so there will be a strong and mutual
understanding between the both. The major gap among
cloud security practice and cloud security research theory
lies in the fact that the assumption in the research go out
some very important differences between actual cloud
security and virtual machine security. Research should be
center on these gaps and differences and its removal. One
of the pieces of the framework might be developing a way
to monitor the cloud's management software, and another
might be development of isolated processing for specific
clients' applications. People' s actions can be followed and
also moni-tored for instance whether people allow the
automated patching software to run, or updating antivirus
software defin-itions, or whether people understand how
to harden their virtual machines in the cloud.
Author proposed a secured architecture with a
multi level authentication technique which will be best
suited for the cloud atmosphere [9]. Service providers can
be use our proposed architecture and technique to preserve
the data security which is now a foremost concern of
individual user/organization. Proposed architecture is
274
taking less res-ponse time to process client' s request and
also decreasing the bandwidth utilization which incurs less
cost. So that is a vast field to enhance our work, future
work may be on decreasing the processing time of the
virtual machines and one can work on other security
issues.
In this paper author worked [10] on multidimensional
password generation algorithm at various authentication
levels has been proposed. By using this algorithm we get
high security for cloud IT framework and also adequate
resource utilization can be done by this proposed
algorithm because here bandwidth utilization is less and
also reduced overall response time. So we find that our
recommended algorithm is enough valuable spare than
already existing techniques on behalf ofmuItidimensional
data security. Furthermore, our future work can be
extended towards the implementation of this algorithm
using CLOUD tools and for better GUI interfacing to
easier communication with client.
Ill. PROPOSED WORK
Client Authentication is a technique by which a
system authenticates and legalizes the user uniqueness of
the system who desires to access it. Client Authentication
guarantees and authorizes the user' s identity through a
program like as a password authentication and guaranties
the genuineness of the application to access it, to make
application effective or valid. Authorization is the
technique of giving users to access their application
according to permissions.
A. Secure Keyword
It is a analog device i.e. exclusively designed to allow
isolation between connected computers. Computers are us-
ually associated to diverse networks and isolation among
these networks must be confident to avoid data leakages
and intrusions. When secure keyboard is use at that time
system keyboard is disabling. And also it is mandatory to
login with secure keyboard when users want to access the
cloud services.
B. Login Process
When the above two process is completed like, the
secure keyboard is connected and you got the access after
that the KBVF procedure is occurred. By few of the
reason the user incapable to verify the above process we
discuss an auto generated alert message is send to the user
registered number.
c. OTP(One Time Password)
OTP is produce spontaneously numeric strings that
authen-ticate the user for a single session on a computer
system or other electronic device. OTP is a static
password then it is not vulnerable to replay attack. While
it will be no longer valid.
Copyright © SMART-2016 ISBN: 978-1-5090-3543-4
 An Efficient Architecture and Algorithm to Prevent Data Leakage in Cloud Computing using Multi-tier Security Approach
D. Image Base Verification System
In image base verification system images are display
in a shuffle way for user to choose our password.
IV. PROPOSED ARCHITECTURE
MSk l I MSk"
MSj l
Fig. 1: Proposed Architecture with Multi-ti er Security
A. User
• User sends a request to the service provider at
master server -1 to access the services.
• Cloud service provider process the request and
sends a registration form to the client.
• User has mandatory to be filled the registration
form along with an active email-id.
• At the time of registration, the users have to give
a numeric number, after registration , which will
be act as a first level security code or pass-word
of the user. Moreover the numeric code will be
static in future use.
B. Master Server-I (MS)
• Master server 1 stores all the registration
information of the clients and it stores the NSC in
encrypted form into the DB.
• At the time of login, when the user entered by
spe-cial keyboard its first level security code,
then it will be encrypted and match with the
stored one. If it is matched, then it proceeds and
sends the required information to the Master
server at level-3 , otherwise it move to Master
server -2.
C. Master Server-I (MS)
• When authentication at the first level of code not
found, Master server at level-2 generate a OTP
security code for second level of authentication and
send it to the registered mobile number of the user.
• The OTP security code has a threshold time, after
this time the code will be invalid.
Copyright © SMART-2016 ISBN: 978-1-5090-3543-4
• After the successful authentication at level-2, if it
is matched, then it proceeds and sends the
required information to the Master server at
level-3.
• Master server at level-2 will take a decision of
nearest server that it can be assign to the user
aimed at provi-ding the cloud services.
D. Master Server-3(MS)
Tier 3
• After successful authentication at the first &
second level of code matching, Master server at
level-3 gen-erate a random image security code
for third level of authentication.
Ti er 2
• After the successful authentication at level-3 , if it
is matched, then it proceeds and user will be
Ti er 1 permitted to access the services.
• Master server at level-3 will take a decision of
near-est server that can be assigned to the user for
providing the cloud services.
E. Authentication
This layer is conscientious for the authentication of
the users. It provides interface to the users to login and
relate with the cloud services provider. Authentication
layer present at master server for checking the authenticity
of the users at multi levels.
F. Cloud Server
The cloud server provides various cloud services like
software, platform, infrastructure (CPU, memory, hard-
ware). Cloud server means virtual server i.e. built hosted
and deliver through cloud computing platform over the
in tern et.
V. PROPOSED ALGORITHM
Multi Iayer authentication (NSC: Alphanumeric
security code, ENSC : Encrypted Numeric Security Code,
RISe:
Random Image Security Code, DB: Database, OTP:
One Time Password, MSI: Master Server-I , MS2: Master
Server-2 MS3: Master Server-3 RMN-Register Mobile
Number)
Step 1: Registration for new users point during the
registration process NSC and also ISC
chooses by the administrator of the
organization an encrypted num-eric code
stored in DB.
Step 2: Numeric User ID and Security code is enter
by special keyboard for the level
authentication and it much match with ENSC
stored in MS 1.
If Match found,
275
5th International Conference on System Modeling & Advancement in Research Trends, 25th _27'h November 2016
College of Computing Sciences & Information Technology, Teerthanker Mahaveer University, Moradabad , India

Then First level authentication is successful VI. PROPOSED TECHNIQUE

and goto step 7.
Cloud Analyst is a tool with a high level of
Else Match not found
visualization ability and based on cloud sim simulator.
OTP has been sent to the RMN.
This is used to measure the overall response time, request
Step 3: OTP at MS2 and sent to the RMN
Step 4: OTP has to be entered by administrator. processing time and the total cost of any task. Here, we
Step 5: Generate new security code at MS2 and sent can set user bases and datacenter in different regions of
to the RMN. the world. User base has many configurable fields like
Step 6: New Password has to be entered for name, region, and requests per user per hour, data size per
authentication. request, peak hours, average users during peak hours, and
If match found then average users during off peak hours. And also we can
Second level is successful con figure the data centers including name, region, nwnber
Step 7: Third level random ISC selected at MS3 by of VM machine and image size, and amount of memory
the administrator of organization stored in and bandwidth availability to single virtual machine. It is
the DB. an evaluation of social networks tool according to
If Match found, then geographic spreading of users and data centers. Cloud
Permitted to user Cloud Services analyst shows a graphical output screen that enables the
Else Match not found users to easily evaluate the result.
Go to step 3
l*I SlInulabOn ~Its

Overall Resp on se Time Summary

Avefage ("1St ~ (ms) Maldn~l (mS)
OIIerallRe~tseT1me: SO.12 39.56 61.63
Oala Cenler Pfoeesslng TIme: 0.50 0.00 • .88

,--- Us.,....
Response Time By Region
-~­ A~(ms) "1tI(ms) .~ M;p(ms) ~
UBI 50117 61631

User Base Hourly Averag. Response nmes

D3ta C.nt.r Request servicing T i mes

r--- Da13 Center ~ 1w9(ms) MItI(ms) M.Cms) -

.5• •
DCI o
0016
~
~
rocr--

-
DC'
-
Fig. 2: Overall Response Time o f First L ayer
r[iJ Simulation Resu lts

Data Canter Hourly Average Processing Times

DC. DC2

0' 2;).3 7 ~ ~ 'O"'2U'.'3'Q171~' ~ :lP2'222? Hr s o , 2 ;) • 3 e 0:; Q ~ 10" 12 IJ ' A 13 IQ 17 18 '~:102122Zl Hr s

I'_·..
Q

I'_""~ 'm. (m.) '~"m.(m.)

Data Center Loading

DC2 ~l'-~
o 1 2 J • S '67 a 9 10 1112 1~ 1~ IS 16 17 la 19 2021 Zl2l H'~

DC. ~r'~'~
01 2 ~ ~ S 6 7 a 9 1 01t121~1~ISI6171aI92021Z12l Hr ~

Fig. 3: Data Center Processing Time o f First L ayer

276 Copyright © SMART-2016 ISBN: 978-1-5090-3543-4

 An Efficient Architecture and Algorithm to Prevent Data Leakage in Cloud Computing using Multi-tier Security Approach

Cost
Total Virtual Mathlne Cost: 11.12
Total Data Transfer Cost: \0.06

Grand Total : $1.79

Data eenter VMCos! Data Transfer Cos! Total

DC4 0.406 0 0.406
DCl 0.406 0 0,406
Del 0.507 0064 0571
Del 0.406 0 0.406

Fig. 4 : TotalCost
rn Simul<'ltion Re-suits

O v e rall Respons e Time Summary

overall Response Time:

Average (ms) Minimum (l11s) Maximum (ms)
50.12 39.56 61.63
I Export Res.JiiSl I
Data Center Processing Time: 0.50 0.00 0.88

Response Time By Region

Userbase Avg (ms) Min(ms) Max(ms)

UB1 50.117 39_558 61.631

User Base Hourly Average Response Times

"" [-.~:" ........... .. - "

Data Center Request SerVicing Ti mes

Data Center Avg (ms) Min(ms) Max(ms)

DC1 o o o
DC' 0.5 0 .016 0_88
DC3
DC'

Fig. 5: O verall Response Time o f Second Layer

~ Simulation R<!:s u lts.

Data Cente,. Hourly Average Process i ng Times

O",ta Center Loading

Fig. 6: Data Center Processing Time o f Second Layer

Copyright © SMART-2016 ISBN: 978-1-5090-3543-4 277

5th International Conference on System Modeling & Advancement in Research Trends, 25th _27'h November, 2016
College of Computing Sciences & Information Technology, Teerthanker Mahaveer University, Moradabad, India

Simulltioli R5ults

Total Vf1ual Machine Cost: 1t72

Total Oala Transfer Cost; '0.06

Grand ToUI : ~1.79

Dala Cenler VM COS! Da!a Transfer Cos! Tolal

DC~ 0106 0 0,406
DCl 0406 0 0.4D6
DC2 0.507 0.064 o.m
DCl 0406 0 0406

Fig. 7: Total Cost of Second Layer

T ABLE I: COMPARISON OF PROPOSED ALGORITHM WITH EXISTING

VII. RESULT COMPARISON
ALGORITHM

Technique Overall Data Center Total Total

3.5 t----1r--"1--------::==--- Response Processing Virtual Data
Time (ms) Time (ms) Machine Transfer
Cost ($) Cost($)
2.5 +----1 Multilevel 150.3 1.46 3.55 0.18
• Data ccntC'f ProccHi ng Authentication
Time(ms) Algorithm
• Virtual Machine (exi sting)
1.5
Proposed 100.24 I 3.44 0.12
• Dam. trnnfcr cost
Algorithm

0.5 We have compared our proposed algorithm with

o existing algorithm. Overall result in our proposed
Multila'eI Authentication PrQPO$cd Algorithm algorithm is better than the existing algorithm.
C'xistins:

VIII. CONCLUSION
Fig. 8: Compari son Graph of Proposed and Exi sting Algorithm
In this paper, we have proposed a secured architecture
160 - , - - - - - - - - - - - - - - - - - with a multi-tier authentication technique which will be
140
best appropriate for the cloud atmosphere. We worked on
client site and server site data security. And in client side
120 we explore the viability of introducing multi-tier
100 authentication to en-sure authentication for cloud access
control. As our proposed tech-nique using three levels of
80 _ ----8-1Jve.-aJ l Response Time
authentication check, authen-tication image check and
60 authentication codes. And authent-ication code are also
stored in the database in encrypted form. To offer more
40
security on client side we use a special key-board. Our
20 proposed architecture and technique are used to sustain the
data security which is now a major issue of indivi-dual
Multi level Authentication Proposed A1 gorithm user/association. On the server site data is stored in
encrypted form on different slave servers. By this, it will
Fig. 9: Comparative Graph of Overall Response Time

278 Copyright © SMART-2016 ISBN: 978-1-5090-3543-4

 An Efficient Architecture and Algorithm to Prevent Data Leakage in Cloud Computing using Multi-tier Security Approach
be hard hitting for the third party hacker to superior the
security codes. In case if hacker hack the one server either
they will not get the security code because of our security
code is stored in di fferent slave server on multilayer. Our
proposed algorithm is adequate enhance valuable than
already existing techniques on behalf of muItilevel data
security in cloud computing.
REFERENC ES
[ I] Dr.AnanthiShesashaayee and D. Sumath, OTP Encrypti on
Techniqu es in Mobiles for Authentication and Transaction Security
International J ournal of Innovative Research in Computer and
Communi cati on Engineering(An ISO 3297: 2007 Certifi ed
Organization)Vol. 2, Issue 10, October 2014
[2] Mohammad F arhatullah and Dr. L.Bullayya Cryptography Based E-
Commerce Security : A Revi ew AL P: A n Authenticati on and Leak
Predi ction M odel for Cloud Computing Privacy IJCS I International
Journal of Computer Sci ence Issues, Vol. 9, Issue 2, No I, March
20121 SSN
[3 ] Hu angshan and P. R. China,Three-Tier Security Model f or E-
Business: Building Tru st and Security for Internet B anking
Services Proceedings of the Second Symposium Internati onal
Computer Science and Computational Technology (ISCSCT ' 09
[4] DeepaPanse, P. Haritha and Keesara, Research Paper Multi-factor
Authentication in Cloud Computing for D ata S torage
SecuritY,lnternati onal J ournal of Advanced Research in Computer
Science and Software Engineering.
[5] SandilyaPemmaraju, V. Sushma and Dr. K. V. Daya.Sagar Data
Leakage Detection u sing C loud Computing G lobal J ournal of
Computer Sci ence and Technology: BCloud and Distributed
Volume 14 Issue 3 Version1.0 Year 2014Type: Double Blind Peer
Revi ewed International Research Journal Publisher: G lobal
Journals Inc. (USA)Online ISSN: 0975-4172 & Print ISSN:
09754350
Copyright © SMART-2016 ISBN: 978-1-5090-3543-4
[6] Chri stian Pri ebe, DivyaM uthukumaran, Dan O' Keeffe David
Eyers,BrianShand,Ru edigerKapitza,Peter Pi etzuchCloudSa-
fetyNet: Detecting Data Leakage between Cloud Ten.
[7] B ernd Zwattendorfer and Dani elSlamani g O n Privacy-Preserving
Ways to Porting the A ustrian elD System t o the Publi c C loud, IFIP
Advances in informati on and communicati on technology volume
405201 3
[8] Prince Jain,Security Issues and their Solution in Cloud
Computinglnternati onal J ournal of Computing & Bu siness
Research ISSN (Online): 2229-6 166 Proceedings of 'I-Society
2012 ' at GKU, Ta
lwandi Sabo Bathinda (Punjab)
[9] AnasuyaSengupta, Brijesh Kumar Chaurasi and Arun Kumar
Yadav, An Effici ent and Secure Architecture for cl oud computing
Environment IEMCON2014 Conference o n El ectronics Engi-
neering and Computer Sci ence
[10] AmbikaG upt,andDr. Arun Kumar Y adav, Multidimensional
Password Generati on algorithm at Multipl e Authenticati on L evels
in Cloud Computing EnvironmentSecond International Conference
on Emerging Research in Computing, Informati on,Communi cati on
and Appli cati ons, August 01-02, 2014
[11] Transl. J. Magn. Japan, vol. 2, pp. 740-741 , August 1987 [Digests
9th Annu al Conf Magnetics Japan, p. 30 I1982].
,
[12] M. Young, The Techni cal Writer's Handbook. Mill Vall ey, CA:
University Science, 1989.
279