See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/307163925

Research on Energy Efficiency in Cloud Computing

Article  in  International Journal of Scientific and Engineering Research · August 2016

CITATIONS READS
4 544

5 authors, including:

Nazmul Hossain Md. Alam Hossain

Jessore University of Science and Technology Jessore University of Science and Technology
13 PUBLICATIONS   15 CITATIONS    25 PUBLICATIONS   61 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Researched on Security Challenges with Possible Solution Strategies in Cloud Computing View project

OAuth-SSO: A Framework to Secure the OAuth-based SSO Service for Packaged Web Applications View project

All content following this page was uploaded by Nazmul Hossain on 28 August 2016.

The user has requested enhancement of the downloaded file.

International Journal of Scientific & Engineering Research, Volume 7, Issue 8, August-2016 358
ISSN 2229-5518

Research on Energy Efficiency in Cloud

Computing
Nazmul Hossain, Md. Alam Hossain, A. K. M. Fayezul Islam, Priyanka Banarjee, Tahira Yasmin

Abstract— Cloud computing is a highly adaptable and efficient infrastructure for running endeavor and web applications. Energy
consumption expenditure and concurrent effects on environment are the dynamic challenges regarding to cloud computing. We
investigated previous researches based on energy efficiency approaches and retrieved the conditions to promote green cloud computing
architectures. Data centers are considered as the backbone of cloud infrastructure. Numerous companies are investing in establishing
large data centers to accommodate different cloud computing services. These data centers absorb huge amount of energy. They are also
very complex in the infrastructure. Over the years, power consumption has emerged as an important factor for measuring computing
resources expenditure. In this paper we will scrutinize all feasible spheres in an ideal cloud infrastructure which are responsible for massive
amount of energy consumption. Here, we will also analyze the methodologies by which energy utilization can be reduced without
compromising Quality of Services and overall performance. In order to make these data centers more energy efficient, many studies have
been proposed. These researches are based on technologies such as consolidation and virtualization. Reducing emissions of carbon
dioxide and energy consumption set up new challenges. The research works for green data centers are derived from the challenges. There
is a current demand of integrated energy efficient cloud framework for establishing data centers that minimizes the impacts on environment
and lessens CO2 emissions. The framework should combine a green IT architecture with different procedures and activities.

Index Terms— Cloud, Green IT, efficiency, cluster, server, data centers, virtualization, PUE metric, GAF, CEC, ASCENT, LEACH.

——————————  ——————————

IJSER
1 INTRODUCTION

T This paper analyzes the energy efficiency of cloud compu-

ting. Since cloud computing is considered as an auspi-
cious technology, it can offer many benefits. Saving ener-
owned systems. Since data centers are used as a basis, in-
volvement of cloud service providers is in the management of
storage devices and other computing resources. So, major part
gy has now a primary preference in almost all sectors of the IT of their investment is in the necessary infrastructure and man-
business. Both corporate users and individual end users are agement. Energy efficiency can be defined as a reduction of
being offered these new services. Services of these types are energy for a given service or level of activity. However, we
generally called cloud computing services .The cloud compu- know about the range and complexity of data center equip-
ting service model comprises of high-capacity storage devices ment. Therefore, it is hard to define unique service or activity
and computing resources that are shared among corporate which could be used to measure its energy efficiency.
and individual users by a service provider.
Nowadays energy efficiency is a thoroughly argued and 2 OVERVIEW OF CLOUD COMPUTING
mostly researched topic in IT sectors. The expanding financial
Cloud computing is generally an internet-based wireless
stress to reduce energy relevant expenditures combined with
computing. Remote servers are connected together to share
the strategy and regulative measures about CO2 emissions are
high-capacity storage devices and computing resources. It
prompting the arrangement and consumption of IT services
acquiesces sharing of data-processing tasks among users by a
faster than ever. Nowadays Green computing is emerging as a
service provider. The combination of computing resources is
critical issue for IT related organizations. Organizations want
managed by various third-party providers. They provide
to be more feasible and liable to offer their resources to the
access to highly efficient computing networks and advanced
governments, customers and to the organization itself. Green
software applications. They also describe a variety of cloud
computing along with green technology is scrutinizing the
computing concepts. They rely on sharing computing re-
environmental friendly use of computers and related re-
sources among users and comprise a large number of comput-
sources. Cloud Computing services are provided by large data
ers where the computers are linked together by a communica-
centers composed of numerous virtualized server instances,
tion network. Cloud Computing is also labeled as "a network-
high- bandwidth networks, as well as supporting systems
oriented computing" [1]. In cloud computing, applications or
such as cooling and power supply. The equipment can be
programs have their ability to perform on several shared com-
formed with hardware and software. Cloud services are ac-
puters at a time. The term "cloud computing" is correlative to
cessed by users through network equipment that connects
distributed computing over a communication network.
servers to the Internet both part of the hardware equipment
and software. User’s software runs on top of servers and is
2.1 Characteristics
handled by Cloud Management System.
There are some important characteristics related to Cloud
Cloud computing basically offers a comprehensive com- computing. It preserves technical issues from the environmen-
mercial benefit, where users (corporate and others) share cen- tal and economic perspective. These characteristics are sup-
trally managed storage devices and other computing re- posed to be the main requirements of the IT industry for fu-
sources. The users do not need to control and operate their
IJSER © 2016
http://www.ijser.org
International Journal of Scientific & Engineering Research, Volume 7, Issue 8, August-2016 359
ISSN 2229-5518
ture improvement. The main characteristics are: Centraliza- 9) Standardization: Standardization has immense impact
tion, service oriented, virtualization, economy of scale, elastici- on cloud adoption and usage. In order to exclude the complex-
ty, dynamicity, Market-Oriented, autonomic, Standardization ity from Cloud, all the operating systems belong to one com-
and Multi-tenancy [2]: pany should be used inside Cloud or one vendor equipment’s
1) Centralization: Centralization means shifting all compu- like vendor routers and switches should be shared among oth-
ting resources which are used for storage, infrastructure and er vendors. Thus it will witness high amount of energy and
applications to the Cloud. So that, we can reduce cost and also gain thrust from different users, vendors and standard bodies.
can have an improved resource management system. 10) Multi-tenancy: Multi-tenancy is one of the key charac-
2) Service-Oriented: We know, everything in the cloud teristic of both public and private clouds. A tenant is known as
means a service. Software, infrastructures, storage and plat- an application. Each tenant needs its own privileged virtual
form are allowed as cloud services that made accessible computing environment. In a typical multi tenancy structure,
through the network. To achieve this, cloud is executed by multiple users do not share each other’s data. But they use the
Service-Oriented Architecture (SOA) model. Moreover, SOA is shared infrastructure. Resources are allocated to users in need.
normally used as an architectural pattern. It is closely coupled, They do not have to concern of location of services and other
allows composition. It is used for building cloud application, users they shared with.
So, Cloud models can be benefited from this architecture in
constructing any of its services i.e. storage, infrastructure etc.
3) Virtualization: The virtualization is accomplished at
platform level. Cloud resources i.e. storage, computing and
networking are virtualized at virtual machine levels. In the
platform level, every application is generalized to one or more
resources provided by different cloud infrastructure provid-
ers. Applications can run within their operating systems on

IJSER
the same physical machine. They can also easily be shifted
from one physical server to another.
4) Economy of Scale: Oftentimes cloud computing is de-
scribed as a technology. But, it is literally an important trans-
formation in the business and economic models for consuming
IT infrastructure that can accelerate a significant cost saving. Fig. 1. Cloud Computing Characteristics.
The capability of cost saving by resource pooling makes the
resource to be shared among the users. Resources can also be 2.2 Cloud Service Models
dynamically allocated according to the applications demand. Cloud computing providers provide their services in ac-
Cloud computing is also an architectural model which is de- cordance with three elementary models. It is generally an as-
signed with multi-tenant applications. So in a cloud compu- sortment of these models/layers to construct the computing
ting model the users have no control over the physical re- structure of the cloud system. Each of them provides different
sources. They are also not aware about the resource location. types of services to the end user. The layers named Infrastruc-
5) Elasticity: Cloud provides the flexibility, dynamically ture as a service (IaaS), Platform as a service (PaaS) correspon-
provision and varies the resources designated for its applica- dingly Software as a service (SaaS). Their characterization be-
tions. The resources (storage, computing, and network capaci- sides their offered services is described below [2]:
ty etc.) will increase or decrease at runtime depending on the 1) Software as a Service (SaaS): SaaS is a software sharing
user QoS requirements. model. In SaaS, a service provider or vendor manages the ap-
6) Dynamicity: It moves virtual storage and machines oc- plications. They make the applications available to the cus-
cupied by the data center. It also helps them to run due to tomers through a network, generally via the Internet. SaaS is
more suitable conditions such as daytime, power consump- enhancing as an increasingly ubiquitous delivery model. It
tion, maintenance concerns and lower cost. Just as the re- appears as an underlying technology that supports service-
sources are also dynamic, they can be changed (increased or oriented architecture (SOA) as well as web services. Through
decreased) depending on the user requirements. the help of SaaS, mature and new instructive approaches are
gaining popularity day by day. SaaS is also usually correlated
7) Market-Oriented: This characteristic dispatches the utili-
with a pay-as-you-go agreement and authorization model.
ty cost dimension of cloud computing. The billing system of
Meanwhile, broadband services have become instantly availa-
cloud computing is likely a utility. So the smallest businesses
ble to provide the user access from more areas around the
can also afford it for their purposes. The users have to pay
world [3].
only for the services that they are using.
SaaS applications must also interact with other data and
8) Autonomic: Cloud services are considered as highly reli-
applications in a symmetrically extended range of platforms
able. They are also autonomic. The autonomic behavior is
and environments. These applications are also proximately
shown by managing themselves in case of the performance
associated to other service models that we have illustrated.
deprivation or failures.
IJSER © 2016
http://www.ijser.org
International Journal of Scientific & Engineering Research, Volume 7, Issue 8, August-2016 360
ISSN 2229-5518
There are two imperceptibly different SaaS delivery models
identified by IDC. It is accepted as a prevalent delivery model
in different business enterprises for their business applications
including small to large businesses that encompasses Manu-
facturing Resources Planning (MRPII), Customer relationship
management (CRM), Human resource management (HRM),
Enterprise resource planning (ERP) along with numerous oth-
er business related applications and platforms [4].
Example: NetSuite, Salesforce.com, Workday, Concur etc.
2) Platform-as-a-Service (PaaS): Platforms are included in
cloud computing for building and successfully executing cus-
tom web-based applications. Platform-as-a-Service is a concept
known for these services. It provides the users with ideal de-
velopment environment to assist the development lifecycle
beginning from the steps: designing, implementation, debug-
ging, testing, deployment. It also supports the operations of Fig. 2. Cloud Service Models.
rich internet application (RIA), a solution stack as a service
and online services. The PaaS model helps to make all the fa- 2.3 Cloud Deployment Models
cilities available required assisting the life cycle of developing
Cloud computing is an archetype that offers different ser-
and delivering various web applications and services wholly
vices in on-demand delivery model [9]. As stated in the pre-
convenient from the Internet. But it does not require down-
viously specified services and computing characteristics, cloud
loading and installing of any software. Therefore, it is greatly
deployment models are classified into four types to work in
helpful for the developers, IT officials, or end users. It is called
consistency with other elements. They are- Public cloud, Pri-

IJSER
the prolongation of the SaaS application delivery model [5].
vate Cloud, Hybrid Cloud and Community Cloud [10].
Example: Google application engine, Aneka, 3Tera, Cog-
1) Public cloud: Public cloud is one of the cloud models
head, Microsoft Azure etc.
that make the services, such as storage and applications avail-
3) Infrastructure-as-a-Service (IaaS): IaaS administrates able for general use through the Internet. Here, the computing
virtualized computing resources through the Internet. Infra- infrastructure is governed by cloud vendors. The customer has
structure as a Service is a software provision model. In IaaS, no physical visibility as well as control over the site where the
an organization redistributes the appliances used to support whole infrastructure is hosted and maintained. The infrastruc-
various networking operations. It includes different hardware tures are shared among different organizations. Service pro-
equipment, servers, storage devices and networking ap- viders host the cloud infrastructure. They made the public
pliances. The appliances are controlled by the service provid- clouds accessible for the general use to the public [11]. Usual-
er. And housing, running and maintaining the appliances is ly, public cloud service providers such as Google, Microsoft,
the responsibility of the service provider. It consists of physi- HP, IBM, Salesforce, Amazon Web Services (AWS) etc. possess
cal or virtual machines, storage and clusters. IAAS layer also and operate the infrastructures. They offer access of their ser-
comprises DBMS systems and other storage relevant services vices over the Internet. Public Cloud is beneficial. Customers
[6]. The infrastructure of IaaS is handled by an upward man- will benefit from this model in economies of scale. Because in
agement layer. It also maintains the runtime environment cus- public cloud, infrastructure costs is shared among all users.
tomization, accounting, application isolation, and QoS services Each singular client is allowed to work on a minimal-cost,
[7]. IaaS virtually leverages the SaaS model that helps to solve “pay-as-you-go” paradigm. Public clouds are ordinarily larger
the identity problem. It also provides for single sign-on for than an in-house enterprise cloud in scale. This provides
web applications, strong authentication process and federation clients with consistent, on-demand scalability.
across boundaries [8].
2) Private Cloud: Private cloud is another cloud infrastruc-
Example: Amazon AWS, Mosso, Joyent etc. ture model that is dedicated to an individual organization.
Private clouds permit organizations to host their applications.
Public cloud infrastructure is lacked of control and data secu-
————————————————
rity. But private cloud concentrates concerns about data secu-
• Nazmul Hossain is currently serving as Lecturer in Computer Science and
rity and control. For security, it is unshared with other organi-
engineering Department in Jessore University of Science & Technology, zations. Generally, public cloud is also controlled internally or
Bangladesh, PH-042172058. E-mail: [email protected] externally by a third-party [12].
• Md. Alam Hossain is currently serving as Chairman in Computer Science 3) Community Cloud: Community cloud is considered as a
and engineering Department in Jessore University of Science & Technology, multi-tenant cloud model. Here the cloud components are
Bangladesh, E-mail: [email protected] shared among particular organizations. Community cloud is
• Others are currently pursuing B. Sc. And M.Sc. degree program in Com- managed, governed and secured generally by all the associat-
puter Science and engineering in Jessore University of Science & Technolo- ing organizations or a third party provider. Community cloud
gy, Bangladesh
IJSER © 2016
http://www.ijser.org
International Journal of Scientific & Engineering Research, Volume 7, Issue 8, August-2016 361
ISSN 2229-5518
concerns about sharing of inner parts of computing infrastruc- the US Congress which concerns about the conventional ener-
ture in between organizations which are systematized in the gy consumption techniques of data centers. Since then many
same community. For example all IT organizations within the governmental companies have started to develop regulations
same state can share the computing resources on the commu- and standards that boost green computing. “Green IT” has
nity cloud to manage data. Community clouds are typically being started to discuss by academia, media and government
hybrid forms of private clouds that are specifically designed to describe about environmental impact on Information Tech-
and managed for a targeted group of users. For this reason, nology (IT). Green cloud computing means the eco-friendly
communities should have identical list of cloud prerequisites. and environmentally reliable use of computers and other re-
Their ultimate objective is to work together to acquire their sources. So, Green Computing can be represented as environ-
business objectives [13].
mentally imperishable computing. It generally addresses the
The purpose of community clouds is to understand the effort/work to expand energy efficiency and power consump-
benefits of a public cloud among participating organizations tion. So that, we can minimizes the cost as well as CO 2 emis-
with extra level of security, privacy and policy acquiescence. sion [15].
So that they can differentiate between the securities usually
In broader terms, the study of designing, manufacturing,
affiliated with private cloud. These clouds can be on-premise
working and adapting computing resources in a way which
otherwise off-premise.
will reduce environmental issues. IT manufacturers and pro-
4) Hybrid Cloud: A formation of two or more clouds which viders are now increasingly investing in the development of
can be public, private or community but remain particular energy efficient computing appliances. They also trying to
entities is called hybrid cloud. These entities are restrained reduce the use of critical materials and encourage the recycla-
together so that they can offer the adjustment of multiple dep- bility of resources. Green computing is renowned as Green
loyment models. The architectural method of hybrid cloud can information technology. The main objective of green compu-
host crucial applications controlled by private clouds. But the ting is to scrutinize new computer systems and applications
applications having comparatively limited security concerns with low cost as well as CO2 emission and increase energy

IJSER
which are controlled by the public cloud can also be hosted by efficiency and lessen power consumption to boost the impe-
hybrid cloud. Thus the management of both public and pri- rishable economic and social development. To make cloud
vate clouds is termed as hybrid cloud. Both on-premise re- greener, we have two ways. First we have to expand energy
sources and off- premise server based cloud infrastructure is efficiency of cloud then supply clean energy. Green computing
required for the architecture of hybrid cloud [14]. is also intent to gain economic sustainability and improve the
computing devices usage policy [16].

Fig. 3. Cloud Deployment Models. Fig. 4. Green Cloud Computing.

2.4 Green Cloud Computing

The omnipresent agitation to acquire economical, ecologi- 3 OVERVIEW OF ENERGY EFFICIENCY
cal and environmental sustainability is arising to reorganize The energy efficiency of an enormous computing infra-
industrial perspective. The prevailing contingency of ecologi- structure can be upgraded at different levels. But identifying
cal deterioration, global warming and the harshness of its po- the power consumed by the components is generally a chal-
tential results clarify the overwhelming demand of environ- lenging task, because it often needs instrumentation. However
mental actions crosswise the world. In 2007, the EPA (Envi- it is crucial to design and evaluate energy-efficient architec-
ronmental Protection Agency) had proposed a statement to tures and algorithms. At overall infrastructure level, such as in
IJSER © 2016
http://www.ijser.org
International Journal of Scientific & Engineering Research, Volume 7, Issue 8, August-2016 362
ISSN 2229-5518
a grid or hybrid acknowledgments can be coordinated to im- tance in energy minimization of network infrastructure. Two
plement energy efficient policies [17]. types of networks are applied in a cloud infrastructure - wired
and wireless network. A report was released by ICT energy
2.1 Energy Efficiency on Cloud Computing where they examines the radio access network and estimates
Instrumentation of an energy efficient cloud infrastructure the energy consumed by the infrastructure. According to re-
requires not only energy efficient host machines. Other ele- port, a large part of the entire energy is consumed by the in-
ments of an entire cloud infrastructure also should be treated frastructure. Moreover the cost aroused for energy consump-
for energy aware appliances. In this part we will discuss the tion is also proportionate with the total expenditure spent for
network maintenance and operations and on personnel. Micro
sections for establishing a typical cloud infrastructure which
Sensor Architectures are now commonly used. It includes four
are responsible for reasonable amount of power dissipation
basic components- power supply, digital processing, radio
[18].
transceiver and sensing circuitry. In this structure, maximum
1) Energy Efficient Hardware: The better way to minimize energy is consumed by radio transceiver while data
the power consumption at host side is the usage of energy processing and sensing circuitry consume very negligible
efficient components at data center, server and virtual ma- amount of energy. The sensor is always remains in one of the
chine level. International standard organizations i.e. US Ener- following four states –receive transmit, sleeping or idle. For
gy Star and European TCO Certification are rating energy effi- energy minimization, sensors should be kept in sleeping state
cient user products. This rating is necessary to estimate the because all other states consume extensive amount of energy.
environmental impact and carbon emission rate of computer GAF (Geographic Adaptive Fidelity) and CEC (Cluster Based
components. Modern electronics equipment like SSD drives is Energy Conservation) and other Topology Control Protocols
designed which is better power efficient than normally used were also propagated for energy minimization in sensor net-
HDD drives. But still they are costly enough [19]. works as well as wireless computing networks. These proto-
2) Energy Efficient Resource Scheduling: Generally, re- cols can identify extravagant nodes and has the ability to turn
them off to conserve energy [22].

IJSER
source scheduling is the essential process for the establishment
of cloud infrastructures. So the necessity of resource schedul- A survey on different techniques used in energy efficiency
ing in a cloud is must. Several analyses have been done which improvement of Cloud systems-
describes regarding resource scheduling process in both vir-
tual machines and computing grid systems. As we know
every server consists of limited resources. In this case, re-
quests/jobs are needed to be scheduled. Applications in cloud
computing is regarded as business processes which includes a
set of conceptual processes. To designate the cloud resources
to their corresponding tasks, we need to first schedule all of
the resources and their corresponding tasks. Till now several
algorithms is been developed about resource scheduling
which deals with availability and reliability of resources. Some
task scheduling algorithms are also being developed. Some are
being developed on the principle of operating system. Many
characteristic including availability, power consumption, re-
sponse time and reliability etc. are needed to be counted while
designing resource schedulers [20].
3) Energy Efficient Clusters of Servers: CPU power dissi-
pation is a major problem for energy efficient servers. Here, Fig. 5. Overview of techniques for improving the energy efficiency of
Cloud Computing systems.
the CPUs absorb the power-driven energy and then this ener-
gy is dissipated by CPU switching devices operations and by
2.2 Data Centre Energy Efficiency
the energy wasted for heating. In energy efficient servers,
power dissipation is basically minimized by optimal CPU uti- Data centers are considered as the biggest energy consum-
lization. Howbeit other cluster components likely memory ers in the cloud infrastructures. Data center consists several IT
devices, storage segments, network peripherals etc. may also equipment, power systems, air conditioning scheme and ener-
consume noticeable amount of power. Therefore a Virtual Ma- gy resources. Energy consumed in a data center classified as
chine may also use significant amount of power-driven ener- site and IT infrastructure. Energy consumed by individual
gy. So, new approaches are designed that aims to reduce this class is relatively equal to each other. Greater part of this ener-
power dissipation in a group at clusters of servers by consider- gy is consumed for powering servers and energy resources
ing system’s throughput and latency [21]. respectively. They consume enormous amount of electrical
energy of a Cloud. Therefore if we can decrease the energy
4) Energy efficient Network Infrastructure in cloud: Re- used by data center even a small amount, we will accelerate an
searchers give importance in reducing energy consumption in energy efficient and empirical Cloud computing [23].
computing elements likely storage, infrastructure, applica-
tions, computation etc. But we do not give as much impor- To minimize the energy usage in a data center is now a hot
IJSER © 2016
http://www.ijser.org
International Journal of Scientific & Engineering Research, Volume 7, Issue 8, August-2016 363
ISSN 2229-5518
issue in the IT industry. Improvement of an efficient data cen- According to different analysis, computing efficiency is mul-
ter doesn’t imply rebuilding the infrastructure and the re- tiplying in every year. Changing or “superseding” outdated
sources within it. Taking some simple steps in energy man- equipment with newest equipment will save massive energy
agement can reduce the expenditures. A data center should be by using the lower idle power consumption and higher effi-
rationalized, virtualized and consolidated. We must review ciency of the newest equipment. However, performance levels
the temperature of the data center day-to-day. Water cooling of computing server are also needed to be increased. Without
method can be considered as an alternative. The volume of a the performance level, we cannot recognize the assurance of
data center should be considered. Renewable energy resources using more efficient hardware.
can be used and the energy distribution systems should be
optimized [24, 25]. 2.4 Energy Efficiency Protocols
The energy efficiency is most important for cloud networks
compared to any other networks. Generally, data transceiver
requires largest energy than the energy required by data
processing in a network [29]. Multiple routes are used to
transmit the data. These routes can build up a communication
between a node and the sink. In order to establish these routes,
the paths composed of multiple nodes that consists maximum
energy resources are selected. Data transmission and battery
power are proportional to each other. Whenever large num-
bers of data are transmitted by the nodes, their battery power
will also proportionately get decreased. In order to minimize
the data size as well as increase the lifetime of battery, aggre-
gation or data fusion techniques can be used. But to get more
efficiency, we generally use energy-efficient protocols. The

IJSER
objective of energy-efficient protocols is to select those paths
that are supposed to maximize the lifetime of network. GAF
and CEC protocols are used to identify redundant nodes and
help the users to turn off them to preserve energy. Other net
Fig. 6. Electrical components of data center. energy efficiency protocols such as ASCENT and LEACH are
also conferred in cloud environment.
2.3 Server Energy Efficiency 1) Geographic Adaptive Fidelity Protocol: GAF is known
Though efficiency of servers is considered as a hot issue in as an energy provision protocol which is actually based on
modern IT infrastructure development but transparency and equivalent node location. In GAF supernumerary nodes are
appliance of efficient server solutions seem to be still limited. determined rely on their geographic locations. In order to bal-
Most of the IT organizations also proclaimed that energy effi- ance the load, the frequency of a node is switched off periodi-
ciency is not still considered as a pertinent principle up to cally. Generally GPS and GAF devices are used to provide the
now. But, awareness regarding energy saving possibilities is location information in GAF appropriately. But it will fail in
actually low. The great news is that energy efficiency technol- applications, if the information about geographic location is
ogy has been currently used in marketing policies. There was not available. For this reason equivalent node concept is used
also the impression is that finally energy efficiency technology by GAF. For communication these nodes are worked as inter-
is now being revolutionized from marketing strategies to prac- mediate nodes because they are equal to each other in terms of
tical usage [26]. their connectivity. In GAF the total network area is separated
into some small virtual grids. So that the nodes situated in
PUE metric is widely used to measure the efficiency of the
neighboring grids are operating in each other’s transmission
server infrastructure including power distribution and energy
range. Thus in GAF any nodes of individual virtual grid can
resources. While PUE metric persist in being a pivotal oppor-
be used for routing and a grid can use multiple nodes when it
tunity for measuring efficiency progression in server infra-
needs to balance the energy usage for that grid. Using GAF we
structure, the potential energy emission reduction rate from
can save energy by keeping only one node working per grid
enhancing PUE will be smaller than the rate accomplished
and turning all other remaining sensor nodes off and keeping
from using cleaner electricity or developing server utilization
them inactive. The nodes are systematically rotated to remain
[27].
operative in a grid operating life cycle. So that only one node
Off-premise coordination facilities can produce more effi- will be active on per grid operating life cycle at a time [30].
cient energy resources management. However, when the serv-
In GAF protocol, the nodes are always switching among
ers are operated at low utilization levels as well as they are
three consecutive states as follows. Initially a node stays in
also using dirty electricity; coordination is only imperceptibly
discovery state with turning its radio on and it can exchange
more advantageous than an on-premise [28].
messages with neighbor node. When an equivalent node per-
Modernizing server appliances to latest advanced models is forms routing, the other nodes either in active state or in dis-
an alternative way to minimize overall energy consumption. covery state can shift to the sleeping state [22].
IJSER © 2016
http://www.ijser.org
International Journal of Scientific & Engineering Research, Volume 7, Issue 8, August-2016 364
ISSN 2229-5518

Fig. 7. State transition in GAF.

2) Cluster Based Energy Conservation Protocol: CEC pro-

tocol is used to measure and maintains the network connectiv-
ity immediately and instantly so that energy can be reserved Fig. 8. Energy Aware Routing Protocol in Cluster based Wireless Sen-
by identifying the node to obtain a typical effective structure. sor Networks.
The nodes are configured by CEC into super-imposed clusters.

IJSER
In CEC, each cluster includes cluster-head together with gate- 3) Adaptive Self-Configuring sEnsor Networks Topology
way nodes. Clustering is an important approach which is used Protocol: ASCENT is considered as a data node selection pro-
to construct flexible sensor networks. The network is arranged tocol rather than a routing or data promulgation protocol. The
into subsets of nodes by a clustering algorithm. In this net- main principle of ASCENT protocol is to simply determine the
work, individual cluster-head node is placed approximately at nodes that should be linked with the routing infrastructure.
the center of the network. So, CEC represents the cluster as a Data promulgation mechanisms like data dissemination or ad-
collection of nodes. In CEC, cluster structure is formed in a hoc routing can run through this multi-hop topology. Hereof,
distributed approach. Nodes are interconnected in cluster routing protocols are considered as subsidiary to ASCENT.
structure. As earlier said each cluster composes a cluster-head, ASCENT protocol is formed of several phases. At first a node
components of the cluster network are in radio range of this initializes itself and then moves into a hearkening-only phase.
node. Each node initially transmits an originating message This phase recognized as neighbor discovery phase. Within
which comprises the node ID together with cluster ID. To per- this phase, individual node acquires an estimation of the
form this operation, the cluster-head is selected first by CEC number of neighboring nodes which actively transmitting
and then it identifies the gateway nodes through which the messages through local partitions. ASCENT does not conduct
clusters are interconnected. When a node recognizes that it with entire network. The network is split into several local
possesses the maximum energy amongst all the neighbor partitions. We understand that enough node density is availa-
nodes, it acknowledges itself as a cluster-head and broadcasts ble to link the partitions and establish the entire region. At
this. Again if a node has the lengthy lifetime amongst its entire first, only a few nodes linked with this network. The other
network neighbor nodes, it is selected as cluster-head. Similar- remain quietly and hearkening to the messages but they are
ly, for selecting the gateway from multiple gateways the ga- not transmitting their messages. Data messages are started to
teway that possesses with lengthy lifetime is assigned the transmit by the source toward the sink. The sink is placed at
highest priority [31]. the range of radio transmission limit. During this phase, an
individual node may join to the network transiently in attempt
The main disadvantage faced with GAF protocol is that
to contribute to upgrade the connectivity. But when a node
global location information is must needed. But we may not
wants to stay for a longer time, it moves into an active phase.
get the information always. With CEC protocol we have the
Within this phase, it starts transmitting routing control mes-
ability to overcome this disadvantage. CEC does not need lo-
sages as well as data messages. Furthermore when a node
cation information for transmission. It determines Network
does not want to join the network, it moves into the adaptive
Redundancy and operates with Distributed Cluster structure
phase. Within this phase, it turns off itself for a period of time.
steps [32].
The following figure shows a simplified design for ASCENT
during initialization [33].

IJSER © 2016
http://www.ijser.org
International Journal of Scientific & Engineering Research, Volume 7, Issue 8, August-2016 365
ISSN 2229-5518
4 CONCLUSION
Cloud computing is expanding rapidly; demands for cloud
implementation is increasing constantly. On the other hand,
cloud providers are trying to host up additional servers to
accommodate anticipated future server requirements. As more
users are joining in cloud and sharing cloud services [36], sig-
nificant changes will be made in economics and greater eco-
nomic scale will be achieved. This will help the cloud provid-
ers to be able to make perfect prediction about cloud capacity
for meeting the demand. Cloud computing has the advantages
that it has enormous potentiality to revolutionize the IT sector
Fig. 9. Simplified schematic of a self-configurable sensor network. by cutting down expenses, improving efficiency and develop-
ing business performance. Thus cloud computing can help us
4) Low-energy adaptive clustering hierarchy Protocol: to achieve a better and a sustainable world. Within its capabili-
Another clustering-based protocol is LEACH. In LEACH, loca- ty, now cloud computing has the ability to reduce more than
lized cluster base stations are rotating randomly in order to 90 percent carbon emissions for most important businesses.
uniformly distribute the energy load between the sensors Future energy savings techniques are likely to evolve as cloud
within the network. Data aggregation strategy is used to re- computing grows.
duce the volume of data to be relayed to the base station. The Here we discussed about energy efficiency at different cloud
advantage of LEACH is that it can gain enough point in reduc- infrastructure and the ways to improve the energy efficiency
tion of energy consumption in contrast with other convention-
of different computing and networking resources [37]. Because
al routing protocol. Conventional protocols regarding LEACH
these resources are believed as the key components of various

IJSER
may be imperfect for static sensor networks. MTE, direct
cloud systems. In view of computing resources, the techniques
transmission, multi-hop routing etc. can be executed with
LEACH [34]. As partitioning a network is time variable, all of energy efficient models are assigned at different architec-
nodes in the network are supposed to be homogenous as well tural levels, including every single node to total infrastruc-
as positive energy-constrained. Network operations are split tures. The techniques are supported by recently advanced
into several rounds. The first phase of each round is called a functionalities and components such as virtualization. Because
set-up step. In this phase, clusters are grouped by small of the analysis of different computing infrastructure, we
transmitting distances. But phase decision is formed by the judged that different energy efficient techniques and ap-
steady-state step independently at the same time. Data trans- proaches should be applied. Some important features alike
ferred to the base station by steady-state step is also compared CPU utilization, QoS, reliability, performance etc. should be
to the setup phase operation. During advertisement step, improved to achieve desired energy efficiency. Additionally,
every individual node decides itself whether it becomes a clus- energy consumption reduction [38] and time complexity re-
ter-head or not [35].
duction should be made highly effective.

REFERENCES
[1] P. Mell and T. Grance, “The nist definition of cloud computing
(draft),” NIST special publication, vol. 800, no. 145, pp. 7, Jan. 2011.
[2] Ole J. Jacobsen, “A Quarterly Technical Publication for Internet and
Intranet Professionals,” The Internet Protocol, vol. 12, no. 3, pp. 6,
Sep. 2009.
[3] R. Buyya, C. S. Yeo, and S. Venugopal, “Market-oriented cloud com-
puting: Vision, hype, and reality for delivering IT services as compu-
ting utilities,” Proc. 10th IEEE Int. Conf. High Performance Computing
Communication, Dalian, China, pp. 5–13, Sep. 2008,
doi:10.1109/HPCC.2008.172.
[4] Kay K. Kim, “The Impact of Operations Manufacturing Management
Systems by Enterprise Resource Planning (ERP) Software Applica-
tion,” International Journal of Science and Engineering, vol. 2, no. 2,
Feb. 2016.
[5] D. Beimborn, T. Miletzki, and S. Wenzel, “Platform as a Service
(PaaS),” Business & Information Systems Engineering, vol. 3, no. 6,
pp. 381–384, Dec. 2011.
[6] W. Voorsluys, R. Buyya, J. Broberg, M. Brock, and Andrzej Goscinski,
Fig. 10. LEACH protocol.
“Cloud Computing: Principles and Paradigms,” John Wiley & Sons,
Inc, pp. 194-196, Mar. 2011.

IJSER © 2016
http://www.ijser.org
International Journal of Scientific & Engineering Research, Volume 7, Issue 8, August-2016 366
ISSN 2229-5518
[7] R. Moreno-Vozmediano, R. Montero, and I. Llorente, “IaaS cloud [23] S. Strutt, C. Kelley, H. Singh and V. Smith, “Data Center Efficiency
architecture: From virtualized datacenters to federated cloud infra- and IT Equipment Reliability at Wider Operating Temperature and
structures,” IEEE Computer Society, vol. 45, no. 12, pp. 65-72, Mar. Humidity Ranges,” White Paper, The Green Grid, USA, Feb. 2015.
2011. [24] “Data Center Power: Managing for Energy Efficiency and Cost Sav-
[8] W. Lloyd, S. Pallickara, O. David, J. Lyon, M. Arabi, and K. Rojas, ings,” White Paper STI-100-013, Quality Rack Power Solutions, Serv-
“Performance implications of multi-tier application deployments on er Technology, USA, Nov. 2013.
infrastructure-as-a-service clouds: Towards performance modeling,” [25] “Energy Logic: Reducing Data Center Energy Consumption by
Future Generation Computer Systems, vol. 29, no. 5, pp. 1254-1264, Creating Savings that Cascade Across Systems,” White Paper, Ex-
Jul. 2013. perts in Business-Critical Continuity, Server Technology, USA, Oct.
[9] I. Buchmann, S. Frischbier, and D. Putz, “Towards an estimation 2010.
model for software maintenance costs,” Proc. 15th European Confe- [26] M. Eastwood, J. C. Pucciarelli, J. S. Bozman and R. Perry, “The Busi-
rence on Software Maintenance and Reengineering (CSMR), pp. 313–316, ness Value of Consolidating on Energy-Efficient Servers: Customer
Mar. 2011, doi:10.1109/CSMR.2011.45. Findings,” White Paper, International Data Corporation, USA, May.
[10] J. Keung and F. Kwok, “Cloud deployment model selection assess- 2009.
[27] J. Subramanian, “Going Beyond Power Usage Effectiveness (PUE) for
ment for smes: Renting or buying a cloud,” Proc. IEEE Fifth Interna-
Data Center Efficiency,” The Data Center Journal, Alpharetta, Geor-
tional Conference on Utility and Cloud Computing (UCC), pp. 21-28,
gia, May. 2012.
Nov. 2012, doi:10.1109/UCC.2012.29.
[28] J. Shuja , S. A. Madani , S. U. Khan , S. Sarwar, K. Bilal and K. Hayat,
[11] W. Jansen and T. Grance, “Guidelines on security and privacy in
“Energy-efficient data centers,” Springer-Verlag GmbH Wien, Vien-
public cloud computing,” NIST special publication, pp. 800–144, Feb.
na, Austria, vol. 94, no. 12, pp 973-994, Dec. 2012.
2011.
[29] Nazmul Hossain and Md. Alam Hossain, “Integrated Cellular and
[12] X. Jing and Z. Jian-jin, “A Brief Survey on the Security Model of
Ad Hoc Relaying Systems: Analysis of Call Blocking Probability &
Cloud Computing,” Proc. 9th International Conf. on Distributed Com- Performance,” International Journal of Scientific & Engineering Re-
puting and Applications to Business Engineering and Science (DCABES), search (IJSER), vol. 6, no. 12, pp 635-643, Dec. 2015.
pp. 475 – 478, Aug. 2010, doi:10.1109/DCABES.2010.103. [30] D. Virmani and S. Jain, “Clustering Based Topology Control Protocol

IJSER
[13] Dina V. Parmar, and Hemina C. Bhavsar, “Implementing Communi- for Data Delivery in Wireless Sensor Networks,” Networking and In-
ty Cloud to Overcome the Problems of Complexity and Security in ternet Architecture, International Journal of Computer Science and
Business Environment,” Indian Journal of Applied Research, vol. 4, Issues, Republic of Mauritius, Jan. 2013.
no. 5, May 2014. [31] S. Parasakthi and P. Mohan Kumar, “EART: Enhancing an Energy
[14] J. Li, Y. Kit Li, X. Chen, Patrick P.C. Lee, and W. Lou, " A Hybrid Aware Routing Protocol in Cluster based Wireless Sensor Net-
Cloud Approach for Secure Authorized Deduplication," IEEE Trans. works,” American Scientific Publishers, California, USA, vol. 13, no.
Parallel and Distributed Systems, vol. 26, no. 5, pp. 1206 - 1216, May 7, pp. 611-617, Jul. 2015.
2015, doi:10.1109/TPDS.2014.2318320. [32] A. Kumar and A. Nayyar, “Energy Efficient Routing Protocols for
[15] A. Jain, M. Mishra, and S. Kumar Peddoju, “Energy Efficient Compu- Wireless Sensor Networks (WSNs) based on Clustering,” Interna-
ting- Green Cloud Computing,” Indian Institute of Technology and tional Journal of Scientific & Engineering Research (IJSER), vol. 5, no.
Nitin Jain National Institute of Technology, pp. 978 - 982, Apr. 2013, 6, pp 440-448, Jun. 2014.
doi:10.1109/ICEETS.2013.6533519. [33] A. Cerpa and D. Estrin, “ASCENT: Adaptive Self-Configuring sEnsor
[16] S. K. Garg and R. Buyya, “Green Cloud computing and Environmen- Networks Topologies,” Technical Report UCLA/CSD-TR 01-0009,
tal Sustainability,” Cloud computing and Distributed Systems UCLA Computer Science Department, Los Angeles, Calif., May.
(CLOUDS) Laboratory, Department of Computer Science and Soft- 2001.
ware Engineering, University of Melbourne, Australia, 2011. [34] F. Xiangning and S. Yulin, “Improvement on LEACH Protocol of
[17] J. J. Green, “Computing Strategies for Improving Energy Efficiency in Wireless Sensor Network,” Proc. International Conference on Sensor
IT Systems,” International Journal of Scientific Engineering and
Technologies and Applications, pp. 260 - 264, Oct. 2007,
Technology, vol. 3, no. 6, Jun 2014.
doi:10.1109/SENSORCOMM.2007.4394931.
[18] T. Mastelic, and I. Brandic, “Recent Trends in Energy Efficient Cloud
[35] F. Xiangning and S. Yulin, “LEACH inspired hierarchical routing
Computing,” Journal of LATEX CLASS FILES, vol. 11, no. 4, Dec
protocols for wireless sensor networks,” International Journal of In-
2012.
formation Technology, Communications and Convergence, vol. 3, no.
[19] U. Wajid, “Energy Efficient Systems, White Paper,” Proc. IEEE SMC
2, Dec. 2015.
2013 conference, Manchester, UK, Nov. 2013. [36] Md. Alam Hossain, Md. Kamrul Islam, Subrata Kumar Das and Md.
[20] Q. Li, and Y. Guo, “Optimization of Resource Scheduling in Cloud Asif Nashiry, “Cryptanalyzing of message digest algorithms MD4
Computing,” Proc. 12th International Symposium on Symbolic and Nu- and MD5,” International Journal on Cryptography and Information
meric Algorithms for Scientific Computing, pp. 315-320, Sep. 2010, Security, vol. 2, no. 1, Mar. 2012.
doi:10.1109/SYNASC.2010.8. [37] A.C. Orgerie, M. D. Assuncao and L. Lefevre, “A survey on tech-
[21] A. Beloglazov, R. Buyya, Y. Choon Lee, and A. Zomaya, “A Taxono- niques for improving the energy efficiency of large-scale distributed
my and Survey of Energy-Efficient Data Centers and Cloud Compu- systems,” International Journal on ACM Computing Surveys
ting Systems,” Technical Report, CLOUDS-TR-2010-3, Cloud Com- (CSUR), vol. 46, no. 4, Apr. 2014.
puting and Distributed Systems Laboratory, University of Mel- [38] J. C. Cano and P. Manzoni, “Evaluating the energy-consumption
bourne, Australia, Jun. 2010. reduction in a MANET by dynamically switching-off network inter-
[22] A. Banerjee, P. Agrawal and N. Ch. S. N. Iyengar, “Energy Efficiency faces,” Sixth IEEE Symposium on Computers and Communications, pp.
Model for Cloud Computing,” International Journal of Energy, In- 186 - 191, Jul. 2001, doi:10.1109/ISCC.2001.935373.
formation and Communications, Science & Engineering Research
Support society, South Korea, vol. 4, no. 6, Dec. 2013.

IJSER © 2016
http://www.ijser.org
International Journal of Scientific & Engineering Research, Volume 7, Issue 8, August-2016 367
ISSN 2229-5518

IJSER

IJSER © 2016
http://www.ijser.org

View publication stats