IOSR Journal of Computer Engineering (IOSR-JCE)

e-ISSN: 2278-0661, p- ISSN: 2278-8727Volume 16, Issue 1, Ver. II (Jan. 2014), PP 66-69
www.iosrjournals.org
www.iosrjournals.org 66 | Page

Improving Cloud Security Using Data Mining

Srishti Sharma
1
, Harshita Mehta
2

1
(BTech Computer Science and Engineering, VIT University Vellore, India)
2
(BTech Computer Science and Engineering, VIT University Vellore, India)

Abstract: Cloud computing is the use of computing resources (hardware and software) that are delivered as a
service over a network (typically the Internet). It does offer great level of flexibility but this advantage comes
with a drawback. With increase in sharing of data over web there is an increase in possibility of data being
subjected to malicious attacks. Attacker/Provider can extract sensitive information by analyzing the client data
over a long period of time. Hence the privacy and security of the users data is compromised. In this paper we
propose an efficient distributed architecture to mitigate the risks.
Keywords: Cloud Computing, Distributed Architecture, Malicious Attacks, Security Breaches, Unauthorized
access

I. INTRODUCTION
A cloud is the integration of cloud computing, networking, storage, management solutions, and
business applications that aid a new era of IT and consumer services. The services provided by cloud are:
infrastructure as a service, platform as a service and software as a service. Cloud services include Infrastructure
as a Service (IaaS), Platform as a Service (PaaS), and Software as a Service (SaaS) [1]. Many of the major
companies like Google, Microsoft and Amazon are providing cloud services. Azure is the cloud computing
service provided by Microsoft that will be used by us for implementation [6]. As we move towards an
interconnected world of numerous clouds, it gives user access to services anytime and anywhere. But there is
also a great risk involved. There are various types of issues that a cloud storage user both at enterprise level and
as an individual consumer might face during the use of the service. Most of the issues are with integrity of the
data, ensuring that the data is confidential and available when it is needed [3]. Also by monitoring the user data
for a long time attacker can extract private information about the user and can use this sensitive data for his/her
personal gain. Thus, confidentiality of the data is compromised [2]. Also the user trusting a cloud provider
might temporarily or permanently lose the data due to malicious attack. One of the major data analyzing and
extracting technique being used now-a-days is data mining. The fact that the entire user data is stored under a
single cloud provider gives an opportunity to the attacker to use powerful mining algorithms and extract crucial
information about the user [7]. In this paper we solve this issue by dividing the data into chunks after
categorizing it and providing these chunks of data to various reliable cloud providers. Categorizing the data
helps in determining the sensitive information. Fragmenting and distributing the data among various providers
minimizes the information stored under one single provider [2]. Thus the attacker will have access to incomplete
information making it difficult for him/her to correlate the data. This optimizes the cost and ensures data privacy
but has a significant performance overhead when user has to access all the data very frequently. We propose to
store the frequently accessed data separately in the respective providers to reduce space and time overhead.
Hence this will prove to be a much more efficient system than the current one.

II. DATA MINING THREATS
Let us consider a scenario where a car Production Company named ABC has trusted a cloud service
provider CLD with critical information that includes the history of tender bidding. Now Mr.X who is a
malicious employee of company ABC, performed linear multiple regression analysis on the data and found out
various dependencies among them. Using these dependencies he calculated the bidding price. Now if X reveals
this information to ABCs rival, ABC might lose the next bidding. But if ABC will distribute this data equally
among three providers then it will become very cumbersome for the attacker to predict the bidding price, thus
saving ABC from a major loss [2].
Suppose a customer shops online from a site. His activity and order summaries are stored in the cloud.
Now if anyone has access to this data he/she can analyze the data and find out dependencies among them. These
dependencies will help in predicting the clients shopping pattern. Thus, distributing the data will combat this
security breach. What we propose is that if the client wants to access his frequent purchases, she/he can easily
access them through cache rather than searching the entire data spread across all cloud providers [10].

Improving cloud security using data mining
www.iosrjournals.org 67 | Page
III. SYSTEM ARCHITECTURE
Cloud Data Distributor and Cloud providers are two major components of the system [1]. The Cloud
Data Distributor will receive data in the form of files from the client which will be split into chunks and
distributed across various Cloud Providers. The Cloud Provider will store the chunk and after analyzing the data
that is frequently accessed by the user he will store this data separately in a different file which will act like the
cache. The cloud provider will respond to the queries of the distributor by providing data from cache rather than
searching through the entire data chunk which consumes time. Hence, this file containing frequently accessed
data will act as a cache memory, increasing the efficiency of the distributed architecture. Clients do not interact
with Cloud Providers directly rather via Cloud Data Distributor. This process can be represented as follows:
Client data Cloud distributor Data split into chunks and assigned to Cloud Providers
Cloud Provider analyzes the data and stores the frequently accessed data in the cache Cloud provider
responds to the queries of the distributor, depending upon the request from client, through Cache.
The system architecture for a single cloud provider is shown in Fig.1.

IV. IMPLEMENTATION
We have implemented the concept of cache memory in a distributed architecture. Distributed
architecture helps in securing users information. Whenever a user supplies data to a cloud, chunks of data are
stored in different cloud providers. When client needs to access all the data frequently, example if client needs
to perform global data analysis on all the data, it includes performance overhead. The client may have to access
data from multiple locations with a degraded performance. To reduce this overhead we analyze the data. The
data is sent to a data mining tool which generates association rules which helps us in determining frequent item
sets(having 100% confidence) using Apriori algorithm ,Carma model etc [5].This is called Mining association
rules in large transactions and relational databases[4]. Intuitively, a set of items that appears in many baskets is
said to be frequent. To be formal, we assume there is a number s, called the support threshold. If I is a set of
items, the support for I is the number of baskets for which I is a subset. We say I is frequent if its support is s or
more [8]. Frequent sets of items from data are often presented as a collection of if-then rules. The form of
association rules is Ij, where I is a set of items and j is an item. The implication of this associated rule is that if
all the items of I appear in some basket, then j is likely to appear in that basket as well. The notion of likely
can be formalized by defining confidence of rule Ij to be the ratio of support for I U {j} to the support for I.
That is, the confidence of the rule is the fraction of the baskets with all of I that also contain j [8].
These are the steps that were followed to implement this concept:
1. The client data is stored in the database in the cloud (Fig.2) [9].
2. This file is then imported to data mining tool (SPSS MODELER) in excel format. Carma modelling was
chosen to generate the association rules and the frequent item sets [11].
3. The data obtained is filtered to obtain the values that have 100% confidence (Fig.3).
4. The data with 100% confidence is then posted on the cloud (Fig.4) [9].

V. FIGURES



Fig.1 system architecture for a single cloud provider

Improving cloud security using data mining
www.iosrjournals.org 68 | Page

Fig.2 client data stored on the cloud


Fig.3 values having 100% confidence are filtered




Fig.4 data with 100% confidence is posted on the cloud

Improving cloud security using data mining
www.iosrjournals.org 69 | Page

VI. Conclusion
Hundred percent secure networks are almost impossible to achieve. New attacks are being discovered
every day and new countermeasures have to be developed to keep data secure. Attackers and providers use
efficient data mining techniques to extract information about the user from the data stored in cloud. A
distributed architecture was proposed to eliminate this threat. But overheads were still prevalent in the system.
Hence cache memory concept was implemented in our system by generating frequent item sets using any data
mining tool.
In the future instead of having a separate cache for every provider we can have a single cache for all
the providers which will store the frequently accessed client data therefore enhancing the efficiency of the
current system.

Acknowledgements
We would like to thank Prof. Geetha Mary.A for her support and guidance.

REFERENCES
White Paper:
[1] Introduction to Cloud Computing Architecture by Sun Microsystems,Inc., June 2009
Journal Papers:
[2] Himel Dev, Tanmoy Sen, Madhusudan Basak and Mohammed Eunus Ali, An Approach to Protect the Privacy of Cloud Data from
Data Mining Based Attacks., sccompanion, pp.1106-1115, 2012 SC Companion
[3] Anup Mathew, The Institute for Computing, Information and Cognitive Systems (ICICS), University of British Columbia,Survey
Paper on Security & Privacy Issues in Cloud Storage Systems, EECE 571B, TERM SURVEY PAPER, APRIL 2012.
[4] Jiawei Han, Data Mining Techniques, SIGMOD '96 Proceedings of the 1996 ACM SIGMOD international conference on
Management of data
Books:
[5] Jiawei Han and Micheline Kamber, Data Mining concepts and Techniques (San Francisco, CA: Morgan Kaufmann, 2006).
[6] William Ryan, Wouter De Kort, Shane Milton , Developing Windows Azure and Web Services, Microsoft Press; 1 edition (14
November 2013)
[7] Siani Pearson, Taking Account of Privacy when Designing Cloud Computing Services, Software Engineering Challenges of Cloud
Computing, 2009. CLOUD '09. ICSE Workshop
Web Sources:
[8] Frequent Itemsets- The Stanford University InfoLab ,http://infolab.stanford.edu/~ullman/mmds/ch6.pdf
[9] Introducing Windows Azure-http://www.windowsazure.com/EN-US/develop/net/fundamentals/intro-to-windows-azure/
[10] J. Salmon, Clouded in uncertainty the legal pitfalls of cloud computing, Computing, 24 Sept 2008.
http://www.computing.co.uk/computing/features/2226701/clouded-uncertainty-4229153
[11] IBM SPSS Modeler Users Guide
,http://faculty.smu.edu/tfomby/eco5385/data/SPSS/SPSS%20Modeler%2015%20Users%20Guide.pdf