An Efficient Data Integration Framework in Cloud Using MapReduce - Computational Intelligence Techniques for Comparative Genomics - page 123

1 Introduction

With increasing amount of digital data, the term big data has been popular nowa-
days; this describes the experimental growth and usage of structured data as well as
semi-structured data. To work with this much of data, there is in need to

nd a new
methodology in a way to produce accurate results [ 1 ]. Nowadays, if we look at the
big data market forecast, it will be grown by 2017 nearly US$50 60 Billions. The

jobs lled in big data are more when compared to other areas. Processing of such
big data in cloud requires an ef

-

cient, scalable, and effective processing and
computational tool such as Hadoop. This open source framework together with
MapReduce has created an evolution in large-scale computing. The four highlighted
features of this are the following: scalability, low cost, ease of usage, and fault
tolerance. Effective solutions were offered by Hadoop to manage and to process this
massive data [ 2 ]. The customization of MapReduce to analyze massive data makes
Hadoop the most preferred and admired technology for handling big data.
Cloud computing has become a hot topic both in research and in industry. It is
any environment which is created in a user s system from an online application on

the cloud and works through a Web browser. Cloud service refers to providing
computing resources through Net remotely. The main cloud computing services are
IAAS, PAAS, and SAAS. Cloud computing allows individual and organizations to
use resources that are managed by service providers at remote locations. The main
characteristics of cloud service are shared resources that are provided dynamically
with pay-as-you-use. When making decisions on deploying or outsourcing this
application into cloud computing-related solutions, security has always been a most
important concern [ 3 , 4 ]. Cloud computing is becoming popular because of its high
reliability, availability, and importantly its low cost; with these

'

exibilities, many
cloud storage services have been deployed. But drawback is performance and
security cannot be guaranteed when data are stored over public cloud.
Data distributors will handle the data to trusted third parties, and hence, there is a
chance of leakage. Some organizations apply information security only in terms of
their network from intruders but with increasing amount of sensitive data as the
growth of the organization leads to increase in number of data points, and some-
times, accidental or even premeditated data leakage from within the organization
has become painful [ 5 , 6 ].

fl

1.1 Contribution

This paper aims to develop non-cryptographic cloud model for cloud computing-
based applications for addressing the authentication and integrity. In this paper, the
non-cryptographic algorithm used was the improved version of Fowler Vo

hash function FNV-1a [ 7 ]. Since it was a non-cryptographic algorithm, it uses only
multiplication and XOR operations repeatedly on the octets of data [ 8 ]. And hence,
the algorithm is fast and uses very less resources.

Noll

- -