Big Data Analytics in the Cloud for Business

Intelligence
K.Venkata Avinash
School of Computer Science And Engineering
The Research And Development Cell
Phagwara,India.

Abstract:

Cloud computing and big data analytics deploying big data analytics through cloud
are, without a doubt, two of the most computing. We argue that cloud
important technologies to enter the computing can support the storage and
mainstream IT industry in recent years. computing requirements of big data
Surprisingly, the two technologies are analytics. We discuss how the
coming together to deliver powerful results consolidation of these two dominant
and benefits for businesses. Cloud technologies can enhance the process of
computing is already changing the way IT big data mining enabling businesses to
services are provided by so called cloud improve decision-making processes. We
companies and how businesses and users also highlight the issues and risks that
interact with IT resources. Big Data is a should be addressed when using a so
data analysis methodology enables by called CLaaS, cloud-based service model.
recent advances in information and
communications technology. However, big Keywords: Cloud Computing, Big
data analysis requires a huge amount of Data Analytics, Cloud Analytics, Security,
computing resources making adoption Privacy, Business Intelligence,
costs of big data technology is not MapReduce, AaaS, CLaaS
affordable for many small to medium
enterprises. In this paper, we outline the
the benefits and challenges involved in
compelled to capture, understand
and harness their data to support
o 1.Introduction : decision making in order to improve
The term Business Intelligence (BI) business operationswas the first
refers to technologies, applications company to properly use the
and practices for the collection, characteristics of cloud computing
integration, analysis, and and to provide their physical
presentation of business resources as virtual resources to the
information. The main purpose of customers [73] followed by others.
Business Intelligence is to support Now different cloud platforms such
better and faster business decision as Google App Engine, Windows
making. Organizations are being Azure, and Salesforce.com etc. are
available in the market, providing
cloud services, which can be and challenges it brings to
utilized by enterprise developers to enterprises. First, we overview the
develop and migrate their concepts, issues and technology of
application and data, and benefit cloud computing and big data
from cloud computing. separately. We then present a
framework that combines these two
In an ever-changing business world, technologies to form an ideal
many companies now face growing platform for e-commerce. We
pressure to develop and ramp up discuss the role of big data in
their business intelligence efforts enhancing the main functional areas
quickly and at a low cost in order to of e-commerce such as customer
remain competitive. Recently management, marketing, payments,
emerged cloud computing is supply chain and management.
changing the way IT services are
provided by companies and how
businesses and users interact with 2.RELATED WORK:
IT resources. It represents a
Cloud computing popularity has prompted
paradigm shift that introduces
several academic and industry initiatives to
flexible service models that
explore the capabilities and enhancements
companies can subscribe on a pay-
in cloud computing. The value proposition
as-you-use model. The data in the
of cloud computing in comparison with on
world is growing exponentially. Big
premise investments is one of the key
data is an evolving term that
research areas. There are several initiatives
describes any huge amount of
to specifically address the security issues
structured, semi-structured and
and challenges in cloud computing. There
unstructured data that has the
have been several academic initiatives
potential to be mined for useful
investigating e-business model aspects of
information. Big data is data that
cloud computing. Aydin discusses research
exceeds the processing capacity of
of E-Commerce Based on Cloud
traditional databases. The data is too
Computing. Dan and Roger compared
big to be processed by a single
various cloud offerings such as Google
machine. The evolving field of big
App Engine, Amazon EC2, and Microsoft
data analytics examines large
Azure to provide guidance on cost,
amounts of data to uncover hidden
application performance (and limitations)
patterns, correlations and other
for different deployment scenarios.
insights. Big data technology has
Agarwal et al present various methods for
become possible with the latest
handling the problems of big data analysis
developments in computer
through Map Reduce framework over
technology as well as algorithms
Hadoop Distributed File System (HDFS).
and approaches developed to handle
In this paper, Map Reduce techniques have
big data. In this paper, our aim is to
been implemented for Big Data analysis
investigate the impacts of cloud
using HDFS. Yadav et al present an
computing and big data on
overview of architecture and algorithms
businesses and analyse the benefits
used in large data sets. These algorithms Standards (NIST). Per the NIST
define various structures and methods definition , “Cloud computing is a
implemented to handle Big Data and this model for enabling ubiquitous,
paper lists various tools that were convenient, on-demand network
developed for analysing them. It also access to a shared pool of
configurable computing resources
describes about the various security issues,
(e.g., networks, servers, storage,
application and trends followed by a large
applications, and services) that can
data set. Fan and Bifet present an overview be rapidly provisioned and released
of big data mining outlining its current with minimal management effort or
status, controversy, and forecast to the service provider interaction. This
future. This paper also covers various cloud model is composed of five
interesting and state-of-the-art topics on essential characteristics, five service
Big Data mining. Sharma and Navdeti models, and four deployment
discuss about the big data security at the models”
environment level along with the probing
of built in protections. It also presents 3.2 Cloud Computing
some security issues that we are dealing Characteristics
with today and proposes security solutions Cloud computing has five essential
and commercially accessible techniques to characteristics. They are on-demand
address the same. The paper also covers all capabilities, broad network access,
the security solutions to secure the Hadoop resource pooling, rapid elasticity
ecosystem. They also provide an overview and measured service. These are the
on big data, its importance in our live and characteristics that distinguish it
some technologies to handle big data. from other computing paradigms.
Jassena and David discuss issues, On-demand Capabilities: A
challenges and solutions of big data consumer can unilaterally provision
mining. Padgavankar and Gupta provide computing capabilities, such as
server time and network storage, as
detail analysis of the challenges involved
needed automatically without
in big data storage and propose some
requiring human interaction with
solutions to handle them. Jayasree each service provider.
provides an overview of big data Broad network access: Capabilities
technologies such as MapReduce and are available over the network and
Hadoop and compares with traditional data accessed through standard
mining techniques. Zulkernine et al mechanisms that promote use by
presents a conceptual architecture for a heterogeneous thin or thick client
cloud based analytics as a service (CLaaS). platforms (e.g., mobile phones,
tablets, laptops and workstations).
Resource Pooling: The provider's
3.Cloud Computing: computing resources are pooled to
serve multiple consumers using a
3.1 Cloud Computing
multi-tenant model, with different
Many researchers have defined physical and virtual resources
cloud computing differently. One dynamically assigned and
mostly accepted definition is given reassigned per consumer demand.
by the United States Institute of Rapid elasticity: Capabilities can
be elastically provisioned and a composition of two or more
released, in some cases distinct cloud infrastructures
automatically, to scale rapidly (private, community, or public)
outward and inward commensurate that remain unique entities, but are
with demand. Measured service: bound together by standardized or
Cloud systems automatically control proprietary technology that enables
and optimize resource use by data and application portability .
leveraging a metering capability at
some level of abstraction
appropriate to the type of service
(e.g., storage, processing,
bandwidth and active user
accounts).
3.3Cloud Deployment
Models Cloud
deployment models are grouped
broadly into four models: private
cloud, public cloud, community 3.4 Cloud Service
cloud and hybrid cloud. Private Delivery Models
cloud is the most secure way to Cloud-based services are grouped
utilize cloud computing. The cloud broadly into four models: Data as a
infrastructure is provisioned for Service (DaaS), Software as a
exclusive use by a single Service (SaaS), Platform as a
organization comprising multiple Service (PaaS), and Infrastructure as
consumers (e.g., business units). It a Service (IaaS). Software as a
may be owned, managed, and Service (SaaS) is a model that
operated by the organization, a third provides the user with access to
party, or some combination of them, already developer applications that
and it may exist on or off premises. are running in the cloud. The access
Community cloud is provisioned for is achieved by cloud clients and the
exclusive use by a specific cloud users do not manage the
community of consumers from infrastructure where the application
organizations that have shared resides, eliminating with this the
concerns. It may be owned, way the need to install and run the
managed, and operated by one or application on the cloud user’s own
more of the organizations in the computers. Platform as a Service
community, a third party, or some (PaaS): is a model that delivers to
combination of them, and it may the user development environment
exist on or off premises. Public services where the user can develop
cloud is provisioned for open use and run in-house built applications.
by the public. It may be owned, The services might include an
managed, and operated by a operating system, a programming
business, academic, or government language execution environment,
organization, or some combination databases and web servers.
of them. It exists on the premises of Infrastructure as a Service (IaaS) is
the cloud provider. Hybrid cloud is a model that provides the user with
 virtual infrastructure, for example
servers and data storage space.
Virtualization plays a major role in
this mode, by allowing IaaS-cloud
providers to supply resources on-
demand extracting them from their
large pools installed in data centres.
Data as a Service (DaaS) is a model
in which, data is readily accessible
through a Cloud-based platform.
Simply put, DaaS is a new way of
accessing business-critical data
within an existing data centre.
Figure 1 illustrates the general cloud
computing architecture.

3.5 Cloud Computing providing in that way continuous

availability of resources. The various cloud
Benefits vendors typically use multiple servers for
maximum redundancy. In case of system
Cost Efficiency - This is the biggest failure, alternative instances are
advantage of cloud computing, achieved automatically spawned on other machines.
by the elimination of the investment in Scalability and Elasticity - Scalability is a
stand-alone software or servers. By built-in feature for cloud deployments.
leveraging cloud’s capabilities, companies Cloud instances are deployed
can save on licensing fees and at the same automatically only when needed and thus,
time eliminate overhead charges such as you pay only for the applications and data
the cost of data storage, software updates, storage you need. Hand in hand, also
management etc. Renting your comes elasticity, since clouds can be
infrastructure can make good financial scaled to meet your changing IT system
sense. The pay as you go (PAYG) model is demands. Fast deployment and ease of
especially Continuous availability - Public integration - A cloud-based application can
clouds offer services that are available be up and running with just a few hours
wherever the end user might be located. rather than weeks or months and without
This approach enables easy access to spending a large sum of money in advance.
information and accommodates the needs This is one of the key benefits of cloud. On
of users in different time zones and the same aspect, the introduction of a new
geographic locations. As a side benefit, user in the system happens
collaboration booms since it is now easier instantaneously, eliminating waiting
than ever to access, view and modify periods.
shared documents and files. Moreover,
service uptime is in most cases guaranteed,
4. BIG DATA ANALYTICS
4.1 What is “Big Data

Big Data is the term for a collection of

data sets so large and complex that it
becomes difficult to process using
conventional data mining techniques and
tools. The overall goal of the big data
analytics is to extract useful information
from a huge data set and transform it into
an understandable structure for further use.
The major processes of big data include
capture, curation, storage, search, sharing, 4.2 Big Data Technologies
transfer, analysis, and visualisation.
Recently the importance of this field has In order to support big data analytics, a
attracted enormous attention because it computing platform should meet the
gives businesses useful information and following 3 criteria, so called 3 Vs as
better insight of both structured and illustrated in Figure 2. Variety: The
unstructured data, which may lead to platform supports wide variety of data and
betterinformed decision-making . In a enables enterprises to manage this data as
business context, big data analytics is the is in its original format, and with extensive
process of examining “big data” sets to transformation tools to convert it to other
uncover hidden patterns, unknown desired formats. Velocity: The platform
correlations, market trends, customer can handle data at any velocity, either low-
preferences and other useful business latency streams, such as sensor or stock
information . Today’s advances in data, or large volumes of batch data.
technology combined with the recent Volume: The platform can handle huge
developments in data analytics algorithms volumes of at-rest or streaming data.
and approaches have made it possible for Traditional data mining involves finding
organisations to take advantage big data interesting patterns from datasets whereas
analytics. Some of the major issues in big data analytics involves large scale
applying big data analytics successfully storage and processing of huge data sets.
include data quality, storage, visualization Traditionally Hadoop and MapReduce are
and processing . Some business examples two of the popular technologies for big
of big data are social media content, data analytics . More tools and
mobile phone details, transactional data, technologies are becoming available for
health records, financial documents, big data processing. Examples include
Internet of things and weather information. Amazon’s Redshift hosted BI data
warehouse, Google’s BigQuery data
analytics service, IBM’s Bluemix cloud
platform and Amazon’s Kinesis data
processing service. The future state of big
data will be a hybrid of on-premises and
cloud, Alternatives to traditional SQL-
based relational databases, called NoSQL
(Not Only SQL) databases, are rapidly
gaining popularity as tools for use in working on huge data has been always a
specific kinds of big data analytic challenge for any trade. Big data has
applications. constructed the road for managing such
huge data making business much simpler
4.3 Big Data Benefits and profitable. Fraud Detection - High-
performance analytics is not just another
The fact that the valuable enterprise data technology fad. It represents a
will reside outside the corporate firewall revolutionary change in the way
raises serious concerns. Some of the most organizations harness data. With new
common challenges are discussed below distributed computing options like in-
Cost reduction - Big data technologies like memory processing on commodity
Hadoop and cloud-based analytics can hardware, businesses can have access to a
provide substantial cost advantages. While flexible and scalable real-time big data
comparisons between big data technology analytics solution at a reasonable cost.
and traditional architectures (data This is sure to change the way insurance
warehouses and marts) are difficult companies manage big data across their
because of differences in functionality, a business – especially in detecting fraud
price comparison alone can suggest order-
of-magnitude improvements. Rather than
processing and storing vast quantities of
new data in a data warehouse, for example,
companies are using Hadoop clusters for
that purpose, and moving data to enterprise
warehouses as needed for production
analytical applications. Faster, better
decision making - Analytics has always
involved attempts to improve decision
making, and big data doesn’t change that.
Following the Big data analytics really
makes the business managers good
decision makers. Large organizations are
seeking both faster and better decisions
with big data, and they’re finding them.
Driven by the speed of Hadoop and in-
memory analytics, several companies are
focused on speeding up existing decisions.
New products and services - Perhaps the
most interesting use of big data analytics is
to create new products and services for
customers. Online companies have done
this for a decade or so, but now
predominantly offline firms are doing it
too. Product recommendation - It is
obviously very clear that the adoption of
big data and analytics have proved to be a
very powerful strategy for online
businesses. The influence of the huge data
of the customers on the business is turning
to be very significant and economic tool
for strengthening a business. Storing and
5. DEPLOYING BIG DATA Data and Information over the net:
Information is available over the network
ANALYTICS IN THE CLOUD and can be accessed anytime through the
Cloud-based big data analytics is a service net by different devices such as laptop,
model in which elements of the big data mobile, ipads etc.
analytics process are provided through a
public or private cloud . It uses a range of Resource pooling: Provider resources are
analytical tools and techniques to help grouped and used efficiently by multi-
businesses extract information from tenant model. Resources include storage,
massive data and present it in a way that is memory, VMs etc.
easily categorised and readily available via
a web browser. Such cloud-based data Rapid elasticity: Resources (both
analytics applications and services are hardware & software) can be increased or
typically offered under a subscription- decreased efficiently and effectively in
based or utility (pay-per-use) pricing quick span of time. Customers can
model. This service model is called Cloud purchase the resources for any quantity
Analytics as a Service (CLAaaS). In this and at any time.
model, analytics is readily accessible
through a cloud computing platform. Such Cost effective: Resource usage can be
cloud-based data analytics service will monitored and would be charged on the
enable businesses to automate processes basis of usage. This system is very
on an anytime, anywhere basis. Examples transparent which makes the provider and
of such cloud-based analytics products and the user more comfortable to adopt it. Big
services include hosted data warehouses, data technologies such as Hadoop and
software-as-a-service business intelligence cloud-based analytics bring significant cost
(SaaS BI) and cloud-based social media advantages when it comes to storing large
analytics. Data stored in a cloud-based amounts of data – plus they can identify
database can help businesses with their more efficient ways of doing business.
decision making processes. With cloud-
based big data, analysts have not only
more data to work with, but also the
processing power to handle large numbers
of records with many attributes. This has
the ability to increase predictability. The
combination of big data and cloud
computing also lets analysts explore new
behavioural data such as websites visited
or location on a daily basis.

5.1 Major Benefits for

Business Organisations

On-demand self-service: As the name

describes, organisations can expand the
storage or service at a click of the button
without any human help. Organisations 5.2 Big Data and Cloud
will can establish big data infrastructure as
quickly as possible. Computing Challenges
 optimized and not limited because of such
policies.

Hacking and various attacks to cloud

infrastructure would affect multiple clients
even if only one site is attacked. These
risks can be mitigated by using security
applications, encrypted file systems, data
loss software, and buying security
hardware to track unusual behaviour
across servers.

Service Delivery and Billing It is difficult

to assess the costs involved due to the on-
demand nature of the services. Budgeting
and assessment of the cost will be very
difficult unless the provider has some good
and comparable benchmarks to offer. The
The fact that the valuable enterprise data service-level agreements (SLAs) of the
will reside outside the corporate firewall provider are not adequate to guarantee the
raises serious concerns. Some of the most availability and scalability. Businesses will
common challenges are discussed below. be reluctant to switch to cloud without a
Data Storage - Storing and analysing strong service quality guarantee.
large volumes of data that is crucial for a
company to work requires a vast and Interoperability and Portability
complex hardware infrastructure. With the Businesses should have the leverage of
continuous growth of data, data storage migrating in and out of the cloud and
device is becoming increasingly more switching providers whenever they want,
important, and many cloud companies and there should be no lock-in period.
pursue big capacity of storage to be Cloud computing services should have the
competitive. capability to integrate smoothly with the
on premise IT.
Data Quality - Accuracy and timely
availability of data is crucial for decision- Reliability and Availability Cloud
making. Big data is only helpful when an providers still lack round-the-clock
information management process is service; this results in frequent outages. It
implemented to guarantee data quality. is important to monitor the service being
provided using internal or third-party
Security and Privacy - Security is one of tools. It is vital to have plans to supervise
the major concerns with big data. To make usage, SLAs, performance, robustness, and
more sense from the big data, business dependency of these services.
organizations would need to start
integrating parts of their sensitive data into Performance and Bandwidth Cost
the bigger data. To do this, companies Businesses can save money on hardware
would need to start establishing security but they must spend more for the
policies which are self-configurable: these bandwidth. This can be a low cost for
policies must leverage existing trust smaller applications but can be
relationships, and promote data and significantly high for the data-intensive
resource sharing within the organizations, applications. Delivering intensive and
while ensuring that data analytics are complex data over the network requires
sufficient bandwidth All these challenges profits and support their decision-making
should not be considered as road blocks in processes. Today it is widely accepted that
the pursuit of cloud computing. It is rather cloud computing and big data technologies
important to consider these issues and the are two dominant technologies that will
possible ways out before adopting the shape up the business world. Cloud is no
technologyGoogle. longer just a buzzword – it’s a fact-of-life
affecting every facet of the technology
industry. Big data technologies provided
through cloud computing will allow
businesses to make proactive, knowledge-
driven decisions as it allows them to have
future trends and behaviours predicted.
Businesses will be able to store their data
remotely and access data and services from
anywhere and anytime. Further, cloud-
based data analytics provides the
infrastructure that companies would
otherwise have to build up themselves
from scratch. Alongside data analytics,
cloud computing is also capable of keeping
businesses stay competitive by providing
many benefits such as cost effectiveness,
resource pooling, on-demand service, rapid
elasticity, and ease of management.
Despite these benefits, there are some
challenges and drawbacks, particularly in
relation to privacy and security. Before
investing in cloud-based big data analytics,
an organisation needs to fully grasp the
extent of what’s involved. Investing in
cloud analytics can be profitable for an
organization but proper planning is
6 CONCLUSIONS essential to ensure that all phases of
analytics elements are covered.
Businesses have long used data analytics
to help direct their strategy to maximise

References :
Cloud Computing for E-Commerce, Journal of Mobile Computing and Application.[ Aydin, N. (2015)].

Talia, D. (2013). Clouds for Scalable Big Data Analytics. Published by IEEE Computer Society.

Fan, J., Han, F. & Liu, H., 2013. Challenges of Big Data Analysis. ResearchGate

Yadav, C. Wang, S. and Kumar M. (2013). “Algorithm and Approaches to handle large Data- A Survey”,
IJCSN
https://nessi.eu/Files/Private/NESSI_WhitePaper_BigData.pdf

http://www.edbt.org/Proceedings/2011-Uppsala/papers/edbt/a50-agrawal.pd

https://ijkie.org/IJKIE_December2014_IRINA%26HAO.pdf