Vous êtes sur la page 1sur 3

International Journal of Emerging Trends & Technology in Computer Science (IJETTCS)

Web Site: www.ijettcs.org Email: editor@ijettcs.org


Volume 4, Issue 2, March-April 2015

ISSN 2278-6856

A Synthesize Virtual Machine Provisioning


Mechanism for Cloud Data Center
Ganesh Kumar Vishwakarma1, Prof. Neelam Sain2 , Prof. Anjul K S Rai3
1

RKDF College Of Engineering,Bhopal, Bhopal-462062,Madhya Pradesh,India

Department of Computer Science & Engineering RKDF College Of Engineering,Bhopal

Head Of Department of Computer Science & Engineering RKDF College Of Engineering,Bhopal

Abstract
Now the days to reduce cost Additional enterprises and
companies utilize cloud environment and services of used and
cost directly proportional time because every service provider
provide service in term as time. In this paper introduce an
algorithm to reduce host selection time that mean give virtual
machine selection time is also reduce. These are help to host
allocation in cloud data center. All service provides want
minimum their action cost over the cloud data centers. This
time blend of Virtual Selection Provisioning Mechanism to
help service provider in simulation. This algorithm used less
execution time CPU Utilization and Minimum Migration
Time in provisioning mechanism. Both mechanism used own
logic at a time for host selection but I used combination of
both algorithm.

Keywords: Virtualization SLA, Eclipse IDE, Cloudsim


3.0

1.INTRODUCTION
Cloud Computing is demonstrating to be a prodigious
technology where computing services are providing above
the computer networks, with on-demand elastic resources
like computing energy, storage bulk, memory and network
[1].cloud service provider have occurred to support IT
revolution and cost lessening for enterprises .For cloud
service providers to deliver various types of cloud services
such as Infrastructure as a Service (IaaS), Platform as a
Service (PaaS) and Software as a Service (SaaS), it is
critical to efficiently preserve their cloud data centers and
reduce superfluous cost on energy consumption. With
thousands of CPUs, memory chips, and networked hard
disks, a cloud data center consumes a lot of electricity to
keep Whole operation mode. When the number of service
users dynamically rises, consistent electricity consumption
of cloud data centers also piles up rapidly.

Figure 1 Layered Cloud Computing Architecture

In this paper we are explain one by one section first is to


introduce of cloud environment and about the paper
section second explain the exiting and relative work,
section third explain the proposed work section 4 is define
the experiment result and graph section 5 explain
conclusion and future work

2.RELATIVE WORK
In [2] explain adaptive heuristics for dynamic VM
consolidation faced some problem and gives some good
algorithm which is overcome problems, but main problem
is 1) Host Overloading Detection and 2)VM selection
21.Mean Absolute Deviation
This policy is used VM allocation policy in cloud data
center.[2] heuristic for deciding the time to migrate VMs
from a host built on utilization thresholds [3]. It is based
on the impression of setting upper and lower utilization
thresholds for hosts and custody the total utilization of the
CPU by all the VMs between these thresholds. If the CPU
utilization of a host sprays below the lower threshold, all
VMs have to be migrated from this host and the host has

Volume 4, Issue 2, March April 2015

Page 206

International Journal of Emerging Trends & Technology in Computer Science (IJETTCS)


Web Site: www.ijettcs.org Email: editor@ijettcs.org
Volume 4, Issue 2, March-April 2015
to be switched to the sleep mode in order to abolish the
idle power consumption. If the utilization overdoes the
upper threshold, some VMs have to be migrated from the
host to reduce the utilization in order to prevent a
potential SLA violation
Some technology to proposed utilization thresholds based
on a statistical analysis of historical data collected in
session of VMs. Robust statistics provides a best approach
to classical statistical methods [4] . . It is a more healthy
estimator of gauge than the model variance or standard
deviation.
For a univariate data set
is
definite as the median of the absolute deviations from the
data value for median:

that is, starting with the residuals (deviations) from the


datas median, the MAD is the median of their absolute
values. We define the upper utilization threshold (u) as
shown in
2.2Minimum Migration Time
This policy to use in VM Selection in cloud data center.
The Minimum Migration Time (MMT) policy migrates a
VM that wants the minimum time to whole a
migration comparatively to the other VMs billed to the
host. The migration time is estimated as the amount of
RAM exploited by the VM separated by the network
bandwidth available for the host . Let j be a set of
VMs currently allocated to the host . The MMT policy
finds a VM that satisfies conditions formalized

where RAMu(a) is the amount of RAM currently utilized


by the VM a; and NETj is the spare network bandwidth
available for the host j.
2.3Minimum utilization
This is goes to minimum utilization of CPU when VM
selection is going on .we are simulate after this policy gets
some effect results to side of service provider. We are help
to maximum no of VM are create in single processor in
single time .that is help also to host releasing in runtime

3.THE
PROPOSED
MECHANISM

PROVISION

Recognized on interconnected nonfiction [2, 3], a


shortened linear relationship amongst power consumption
and CPU exploitation is embraced as the power model for
cloud data center Mostly, VM migration is considered as
systematic migration and animate migration. Systematic
VM migration exchanges a VM by stopping the VM
inside the current host, replication equivalent memory
substances of this VM to some impermanent buffer or
hard disk, resurrecting the VM on the novel host and then
replenishing the memory substances backbone to the VM

Volume 4, Issue 2, March April 2015

ISSN 2278-6856

from the temporary buffer. Live VM migration achieves


the identical coherent migration procedure while it does
not essential to halt the VM through the evolution period.
We adopt alive migration manner is embraced for VM
provisioning. The migration charge model is expressed as
the presentation poverty proportion (i.e. 10% based on [2])
times the CPU exploitation ratio of the embattled VM
through its conscious migration time period. The
conscious migration interval is premeditated as the worth
of the whole quantity of memory used by the VM
separated by accessible network bandwidth. Operator SLA
is demarcated as the whole computing resource entreated
by a VM, which is signified by the quantity of Million
Instructions Per Second (MIPS). If obtainable MIPS
quantity from a host cannot fulfill the MIPS requirements
from VMs which exist in on it, we request that a SLA
destruction transpires. Our mechanism select one policy
which is best our host system MadMmt /MadMu and
select one and continue process will be same
The host overloading recognition scheme, the VM
selection outline and the VM appointment outline. The
host overloading recognition outline governs when to
migrate VMs from one host to another host. The VM
selection outline regulates which VM(s) to be migrated
from an overloading host. The VM settlement outline
regulates to which host(s) the migrating VM(s) should be
located. The CPU consumption ratio of a host is
demarcated as the worth of the entire assigned MIPS for
VMs separated by the whole MIPS volume in the host. If
the CPU consumption ratio of the embattled host does not
permit the pre-defined beginning, then this outline will
checked whether the existing number of VMs in the host
is more than the whole number of VMs exists in at the
identical host when SLA violation transpired last time. If
either one of the two beginnings has been passed, then the
embattled host arrives the overloading cautionary status.
The VM selection outline will be initiated once a host is
incoming the overloading cautionary status. Minimum
Migration Time (MMT) policy [2] is recognized as the
VM selection outline to recognize and select a VM that
requirements the minimum time to achieve its migration
transition in assessment with the other VMs exist in at the
similar host. Minimum Utilization policy is recognized as
the minimum cpu used and select a VM requirement the
minimum cpu utilization in same host is exits . To
residence the recognized VM(s), this policy provide
excellent performance

Algorithm 1 VM Selection Policy


1. Input : HOST Output: Vm ID to Migrate
2. getMigratableVms(host) PowerVmList
3. if migratableVms.isEmpty()then return NULL;
4. assign Vm vmToMigrate = null;
5. assign minMetric = Double.MAX_VALUE ;
6. foreach VM in migratableVms then
7. if vm.isInMigration() then continue;
8. assign metric = vm.getRam();

Page 207

International Journal of Emerging Trends & Technology in Computer Science (IJETTCS)


Web Site: www.ijettcs.org Email: editor@ijettcs.org
Volume 4, Issue 2, March-April 2015
9. assign
metric1
=
vm.getTotalUtilizationOfCpuMips(CloudSim.clock())
/ vm.getMips();
10. if metric < minMetric && metric < metric1 then
11.
minMetric = metric ;
12.
vmToMigrate = vm;
13. Elseif metric1 < minMetric then
14.
minMetric = metric1 ;
15.
vmToMigrate = vm;
16. return vmToMigrate;

4.SIMULATION EXPERIMENTS
CloudSim 3.0 is used to build the Recommended VM
Selection provisioning mechanism and preform simulation
experiments. In order to simulate our mechanism in cloud
data center in IDE .we utilize real data of power
consumption delivered by SPECpower benchmark. We are
simulate in some host machine. We are takes Physical
host machine is HP Proliant ML 110 G4. Inside machine
xeon OS ,2*(1860 MHZ cpu with 2 core) and 4 GB
RAM.and Simulation machine is Dell .inside machine
window 8 OS ,2.40 GHZ cps with 2 core and 6 GB
RAM
We gets Result as comparison exits provisioning
mechanism in VM selection policy Mad Minimum
Migration Time and Minimum Utilization
Table 4 Comparison of policy

Figure 2 graph between policy


1. Minimum Migration Time VM selection policy
2. Minimum Utilization VM selection policy
3. Our mechanism

Volume 4, Issue 2, March April 2015

ISSN 2278-6856

5.CONCLUSION AND FUTURE WORK


This paper elaborates the concept of virtual machine
selection and provisioning playing prominent role in
assigning the tasks in a cloud computing environment for
equitable load distribution. Improve response time of jobs
and removing the situation of node overloading and under
loading in the system. Using over algorithm. we get less
execution time over VM selection policy. Our algorithm
provide less virtual machine migration to increase the
storage capacity of both CPU and RAM. In future work,
improve the VM selection policy to take more and more
advantages to update the storage capacity of sever system
and host system and get much better performance to
favors of customer

REFERENCES
[1]. VUPIC Virtual Machine Usage Based Placement in
IaaS Cloud Gaurav Somani, member, IEEE, Prateek
Khandelwal, and Kapil Phatnani
[2]. Anton Beloglazov, and Rajkumar Buyya, "Optimal
Online Deterministic Algorithms and Adaptive
Heuristics for Energy and Performance Efficient
Dynamic Consolidation of Virtual Machines in Cloud
Data Centers", Concurrency and Computation:
Practice and Experience (CCPE), Volume 24,Issue
13, Pages: 1397-1420, John Wiley & Sons, Ltd, New
York, USA, 2012
[3]. Beloglazov A, Abawajy J, Buyya R. Energy-aware
resource
allocation
heuristics
for
efficient
management of data centers for cloud computing.
Future
Generation
Computer
Systems2011
doi:10.1016/j.future.2011.04.017
[4]. Huber PJ, Ronchetti E, Corporation E. Robust
statistics, vol. 1. Wiley Online Library, 1981.
[5]. Kumar S, Talwar V, Kumar V, Ranganathan P,
Schwan K. vManage: loosely coupled platform and
virtualization management in data centers.
Proceedings of the 6th international conference on
Autonomic computing (ICAC 2009),Barcelona,
Spain, 2009; 127136.
[6]. Barford P, Crovella M. Generating representative web
workloads for network and server performance
evaluation. ACM SIGMETRICS Performance
Evaluation Review 1998; 26(1):151160.
[7]. Cleveland WS. Robust locally weighted regression
and smoothing scatterplots. Journal of the American
statistical association 1979; 74(368):829836.
[8]. Shi, Y., Jiang, X., & Ye, K. (2011, September). An
energy-efficient scheme for cloud resource
provisioning based on cloudsim. In Cluster
Computing (CLUSTER), 2011 IEEE International
Conference on (pp.595-599). IEEE.
[9]. Li, X., Jiang, X., Ye, K., & Huang, P. (2013, June).
DartCSim+: Enhanced CloudSim with the Power and
Network Models Integrated. In Cloud Computing
(CLOUD), 2013 IEEE Sixth International Conference
on (pp. 644-651). IEEE.
Page 208

Vous aimerez peut-être aussi