Energy Saving by Virtual Machine Migration in Green Cloud Computing

International Journal of Trend in Scientific Research and Development, Volume 1(1), ISSN: 2456-6470
www.ijtsrd.com
46
IJTSRD | Nov-Dec 2016
Available Online@www.ijtsrd.com
Energy Saving by Virtual Machine Migration in
Green Cloud Computing
Ruhi D. Viroja
M.Tech, CSE student, Institute of Technology,
Nirma University, Ahmedabad, Gujarat, India
Dharmendra H. Viroja
Asst. Prof., Physics Dept., Arts, Commerce &
Science College, Borsad, Gujarat, India
ABSTRACT
Nowadays the innovations have turned out to be so
quick and advanced that enormous all big enterprises
have to go for cloud. Cloud provides wide range of
services, from high performance computing to
storage. Datacenter consisting of servers, network,
wires, cooling systems etc. is very important part of
cloud as it carries various business information onto
the servers. Cloud computing is widely used for large
data centers but it causes very serious issues to
environment such as heat emission, heavy
consumption of energy, release of toxic gases like
methane, nitrous oxide, carbon dioxide, etc. High
energy consumption leads to high operational cost as
well as low profit. So we required Green cloud
computing, which very environment friendly and
energy efficient version of the cloud computing. In
this paper the major issues related to cloud computing
is discussed. And the various techniques used to
minimize the power consumption are also discussed.
Keywords: Cloud computing, green cloud computing,
virtualization, resource allocation, VM migration etc
1. INTRODUCTION
The increasing demand of large processing is handled
by data centers by providing sharing computations
and system resources to fulfill services over the
internet. Cloud is a collection of virtual computer
resources and very high capacity storage devices.
Various third-party service providers manage the data
sharing of these remote resources among the end
users. They provide access to the high speed network
and resources by providing variety of cloud services.
As the large number of computers are linked together
via shared network to create cloud, cloud computing
is also called “a network oriented computing” [1].
Various services provides by cloud:
i. Infrastructure as a service (IaaS): This provides
use and pay facility to users. The system
resources are shared to users by various vendors.
The client can manages OS, system storages,
network connection, application software etc.
ii. Software as a service (SaaS): Based on the
user/client requirement, the available software on
the cloud is given to that client. The clients pay
for services as they use.
iii. Platform as a service (PaaS): It allows the clients
to use the platform so they can store their own or
personal software as well as applications on the
subscribed cloud [7].
Fig 1- Cloud services [3]
2. Requirement of green cloud computing
With the expanding interest of online administrations
in each territory, the need of distributed computing is

International Journal of Trend in Scientific Research and Development (IJTSRD) ISSN: 2456-6470
47
expanding quickly. With this expansion, the need of
more servers increments. Along with this it also
increases energy requirements. The measurement
shows that an ideal server makes use of around 70%
of its energy which is total waste [8].
Hence it turns out to use the energy in efficient
manner. The green processing or computing utilizes
the administrations of distributed cloud computing in
proficient way. It uses the system resources in an eco-
accommodating and proficient way. In the present
pattern in IT businesses, going green has turned into a
motivation for advertising and to lessen the various
costs. One of the best approaches to accomplish the
energy efficient model is the Virtualization.
3. Problems related to cloud computing and need
of green cloud computing
With the fast going life, everyone is using computer
applications and different IT services. This leads in
more power consumption and ends into emission of
toxic gases like carbon dioxide and into the
environment. The data centers needs more power for
processing and also consume power even they are
idle. If the proper cooling mechanism is not used then
due to heat consumption it will lead to loss of whole
energy. Due to adoption of cloud computing more
issues arise which leads towards the use of green
cloud computing. Main considerations which are in
charge of the move from cloud to green cloud are:
3.1 More power consumption
To provide high processing power, the largescale
datacenter needs very large amount of energy. This
energy consumes big portion of the total cost of
operation [9]. According to the research done by
Gartner, this energy occupied 10% of total cost which
may be increased to 50% in upcoming future. Another
research is also done that an idle datacenter occupies
nearly 70% of its peek energy which is also very
sensitive matter as it decrease the efficiency of the
datacenters. As per Gartner report, an average
datacenter occupies very large amount of power
which is more than enough to serve nearly 25000
homes as a power source [10].
3.2 Generation of heat
It is extremely important to decrease the measure of
energy required to create and manage the cloud. Large
amount of power consumption generates heat which
should be maintained at minimum level by applying
some cooling techniques to the datacenters.
Approximately 70% of produced heat is because of
large scale infrastructure used to build cloud. Any
failure regarding maintenance of datacenters affect the
reliability of the datacenter which may break SLA
between cloud vendor and customers [11].
3.3 Emission of toxic gases
The excess processing at datacenters also creates
problem regarding emitting out vary toxic and
harmful gases like carbon dioxide and another carbon
compounds which leads to global warning. So this
should be minimized to keep our environment neat
and clean.
3.4 Security
If any technology is scalable enough then only it is
useful else it will break at some point as it cannot
fulfill users’ incrementing demands. Along with
scalability it also provides integrity and security.
Cloud computing require resource management
through proper connection/network channel between
the users. Using virtual machine concept, multiple
users of same machine can share network. This
requires highly secure channel otherwise it will affect
the integrity of the data being transmitted on that
channel.
Due to above and many more other factors, the
present cloud concept should be modified. So Green
computing is a solution with which we can create
environment-friendly as well as energy efficient
computing system.
4. Green computing
The term "Green figuring" implies practicing
environmental safety with computer frameworks. This
mainly focuses on creating various peripherals,
processors and servers of the datacenters energy
efficient by using minimum system resources and
doing optimum e-waste management. So design,
development, usage and waste management of the
entire physical component etc. are very environment-
friendly and highly efficient with green computing
[6].
It attempts to minimize the power consumption and
carbon dioxide emission. So we can say that green
computing aims to create whole computer system

48
with different software services at very low cost and
low power consumption to maintain eco-friendly
environment.
5. Strategies and solutions to reduce energy usage :
Making cloud more green
To make the distributed cloud computing more
ecology friendly the examinations have been done on
principally three methodologies. The approaches that
are applied to the datacenters are [12]:
5.1 Dynamic Voltage Frequency Scaling (DVFS)
Every last electronic circuit works on their clock
signal. By adjusting this clock frequency we can
manage the input voltage of the circuit but it considers
the characteristics of the hardware. It cannot handle
the dynamic nature of requirement so power saving
will be less compared to other techniques. Another
reason of being less popular is that it is only applied
to CPU, not applied to another component of the
system. To reduce the total power consumption the
idle servers can switched to sleep mode.
5.2 Algorithms
It is practically proved that an idle datacenter occupies
nearly 70% of its peek energy. The prediction using
neural network and green scheduling algorithms
identifies the approximate dynamic workload of the
server. So we can shut down the unoccupied servers
to minimize the total energy usage. To fulfill the SLA
sometimes some extra servers need to be added. So
the green algorithms should be capable enough to
minimize cost, power consumption and should be
environment-friendly by ensuring quality of services.
5.3 Dynamic resource allocation
Various virtual machines can be handling by physical
machine upon which different applications can be run.
In cloud, for load balancing these virtual machines are
transferred to another host with lower load. This can
happen because of dynamic nature of user
requirements or non-availability of system resources.
While transferring/migrating the virtual machine
(VM) the care should be taken that it won’t lead to
more power consumption. So VMs are migrated
among the processing nodes which are power
efficient. This method is discussed in detail latter.
6. Proposed approach
The datacenters located at different location has
unique carbon footprint rate according to various
energy resources. Following equation gives carbon
footprint for particular cloud with d datacenters which
consists of c clusters having total n hosts and time
period is [0,t] :
𝐶𝐹 =
(𝑃𝑈𝐸(𝑖) × (𝑐𝑓(𝑗) ×
𝑘=1ℎ 𝑃𝑣𝑚𝑖,𝑗,𝑘,𝑡×ℎ 𝑡))
(1)
Here,
CF = carbon footprint of the cloud
PUE = power usage effectiveness and PUE is the ratio
of total power consumption to that by IT devices.
ht = holding time for virtual machine (vm)[4].
When there is a requirement for any resource it is
forwarded to host having enough RAM, CPU,
processing elements as well as storage. This
procedure can be done in two simple steps:
Step-1:
Broker of the cloud maintains the repository of
datacenters containing details about carbon footprint
rate. Receiving upon any request the broker will
transferred it to proper host having minimum rate
after visiting the repository.
The algorithm to allocate VM to the carbon efficient
datacenter is given bellow:
1. Request for VM
2. Retrieve information about datacenters from
repository
3. Retrieve all available hosts in the cloud
4. Make a list of hosts fulfilling VM requirement
5. Select most efficient host and allocate the host
to VM
Step-2:
The actual migration takes place in second step.
Depending upon the present state of the host, it is
migrated to such host having underutilization of
workload. When a certain host crosses upper

49
threshold value of CPU utilization, some VMs are
migrated to another host and if host goes
underutilization then all VMs are migrated to another
host leaving underutilized host in sleep mode. The
minimization of VM migration approach [5] helps to
select minimum number of VM to be transferred.
The selection of VM is carried out based on following
formula:
R =
⎩
⎪
⎨
⎪
⎧
𝑆|𝑆 ∈ 𝑃(𝑉𝑗), 𝑈𝑗 −
∑ 𝑈𝑎(𝑣) < 𝑇𝑢∈
,
∑
( )
( )∈ → 𝑚𝑖𝑛
𝑉𝑗, 𝑖𝑓 𝑈𝑗 < 𝑇𝑙;
∅, 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒
, 𝑖𝑓 𝑈𝑗 > 𝑇𝑢;
(2)
Where
Tu = upper threshold of utilization
Tl = lower threshold of utilization
Vj = VMs at jth
host
Uj = CPU utilization of jth
host
Ua (v) = CPU utilization of VM v
Ur (v) = CPU capacity requested for VM v
The modified best fit decreasing (MBFD) algorithm
makes the list of all VMs by sorting them based on
their CPU usage in descending order. This algorithm
makes mapping between the hosts and VMs such that
it reduces the power consumption. The selected VMs
which are required to migrate are placed by using
MBFD algorithm [5].
7. Conclusion
In this paper basics of cloud computing, its current
problems and how it can be eliminated by introducing
the concept of green cloud computing is discussed.
The proposed technique for allocation of resources
and migration of VM in datacenters reviews the
carbon emission rate of the datacenters. So far a
minimal work is done to minimized power
consumption keeping system efficient, reliable, secure
etc. and more attempts are required in this critical
issue as it affects the environment so our life too.
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) Shivangi Sharma, Gaurav Sharma, ”A Review on
Secure and Energy Efficient Approaches for
Green Computing”, International Journal of
Computer Applications (0975 – 8887) Volume
138 – No.11, March 2016.
3) Sushil kumar Choudhary, R.S Jadoun,H.L
Mandoriya, Ashok Kumar,” Latest development
of cloud computing technology, characteristics,
challenge, services & applications” ,IOSR Journal
of Computer Engineering (IOSR-JCE)e-ISSN:
2278-0661,p-ISSN: 2278-8727, Volume 16, Issue
6, Ver. V (Nov –Dec. 2014
4) Minyi Yue, "A simple proof of the inequality FFD
(L)<11/9 OPT(L)+1,for all l for the FFD bin-
packing algorithm", in Acta Mathematicae
Applicatae Sinica (English Series), Oct 1991, vol.
7,issue 4, pp. 321–331.
5) Anton Beloglazov, Jemal Abawajy, Rajkumar
Buyya, " Energy-aware resource allocation
heuristics for efficient management of datacenters
for Cloud computing", in Future Generation
Computer Systems, 2012, pp. 755-768
6) Pushtikant Malviya and Shailendra Singh,‟ A
Study about Green Computing”, International
Journal of Advanced Research in Computer
Science and Software Engineering, vol. 3, Issue
6, pp. 790-794, June 2013,
http://www.ijarcsse.com/docs/papers/Volume_3/6
_June2013/V3I6-0320.pdf
7) Anil Kumar Raghuwanshi , “ Capacity
Management for Virtualized Data Centers using
ECIES and Scheduling”, International Journal of
Computer Applications (0975 – 8887) Volume
103 – No.10, October 2014
8) Pallavi Mohindru, Amanpreet Kaur, “A Review
on: Energy Efficiency in Green Computing”,
International Journal of Advanced Research in
Computer Science and Software Engineering,
Volume 5, Issue 4, 2015
9) Jayant Baliga, Robert W. A. Ayre, Kerry Hinton,
and Rodney S. Tucker, „Green Cloud Computing:
Balancing Energy in Processing, Storage, and
Transport‟ , Proceedings of the IEEE, vol.1,

50
January 2011, pp. 149-167DOI:
10.1109/JPROC.2010.2060451
10) Anuj Prasher and Rajkumari Bhatia,” A Review
On Energy Efficient Cloud Computing
Algorithms”, International Journal of Application
or Innovation in Engineering & Management
(IJAIEM), vol. 3, Issue 4, April 2014
11) Dzmitry Kliazovich, Pascal Bouvry and Samee
Ullah Khan”, GreenCloud: a packet-level
simulator of energyaware cloud computing data
centers,” Springer Science+Business Media, LLC
2010,
http://link.springer.com/article/10.1007%2Fs1122
7-010-0504-1
12) Kalange Pooja R,” Applications of Green Cloud
Computing in Energy Efficiency and
Environmental Sustainability”, IOSR Journal of
Computer Engineering (IOSR-JCE) ISSN: 2278-
0661, ISBN: 2278-8727, PP: 25-33
www.iosrjournals.org

Energy Saving by Virtual Machine Migration in Green Cloud Computing

More Related Content

What's hot (19)

Similar to Energy Saving by Virtual Machine Migration in Green Cloud Computing (20)

More from ijtsrd (20)

Recently uploaded (20)

Energy Saving by Virtual Machine Migration in Green Cloud Computing