2. The main motive of High Performance Computing
(HPC) is to enhance the computing power of a
computer system. To implement HPC, various
methods are introduced, as
Cluster Computing
Grid Computing
Cloud Computing
Providing a HPC environment is one of the
primary goals of the Cloud Computing. Cloud is
a model in which resources are scattered in
distributed manner; there is no any central
authority to control the resources. Cloud
computing environment is categorized on the
basis of scale and functionality. Cloud computing
is a pay-per-use mechanism in which a user have
to pay if he/she access the resource of any other
node.
INTRODUCTION
3. Cloud Computing:
NIST definition of cloud computing- Cloud
computing is a model for enabling ubiquitous,
convenient, on-demand network access to a
shared pool of configurable computing
resources (e.g., networks, servers, storage,
applications, and services) that can be rapidly
provisioned and released with minimal
management effort or service provider inter-
action .
Load Balancing And Virtualization:
In cloud computing, load balancing basically
means adjusting the loads across the nodes
forming the cloud which may be the CPUs,
network links or other resources.
Two types of virtualization are:
1. Full Virtualization
2. Para Virtualization
LITERATURE REVIEW
4. INTRODUCTION
The main motive of High
Performance Computing (HPC)
is to enhance the computing
power of a computer system. To
implement HPC, various
methods are introduced, as
Cluster Computing
Grid Computing
Cloud Computing
5. Load Balancing Algorithms:
Algorithms can be categories into two major categories: static or dynamic.
Static Load Balancing Algorithm:
As shown in Figure 1, Static load balancing algorithms allocate the tasks of
a parallel program to workstations based on either the load at the time
nodes are allocated to some task, or based on an average load of our
workstation cluster.
• Resource
Information
Process
• Application
Process
• Scheduler
Process
• Pre
Knowledge
Base
Process
6. Dynamic Load Balancing Algorithm:
As shown in Figure 2 Dynamic load balancing algorithms make changes to
the distribution of work among workstations at run-time; they use current
or recent load information when making distribution decisions.
Parameters for Load Balancing:
Performance of a grid is not affected by a single parameter. Dynamic load
balancing in grid computing is done on the basis of different parameters. If
we take a combination of different parameters to decide load balancing, it
may give higher performance than taking a single parameter [11].
Different parameters for load balancing are given below:
7. Network Parameters:
There are some communication network parameters that can affect the load
balancing in grid computing environment. These are: available network
bandwidth, communication link capacity, inter-site communication delay
and network latency [12]. In our work, we are considering inter-
communication delay which provides better performance than other
network parameters.
Communication Node Characteristics:
Characteristics of a node mean the availability of required resources on the
node. Communication node characteristics can affect performance by three
parameters: number of available processing units, processing speed and
memory capacity [13]. Here we are considering number of available
processing element for decision making.
Application characteristics:
The application transferred to other computing node should be compatible
with transformation of job between or among the computation node. There
are some parameters related to the application which affect the load
balancing are pre-emptive applications, non-preemptive application and
remaining execution time [14]. Remaining execution time can get higher
performance than other parameters. We are considering execution time for
automatic selection process.
8. Analytical Hierarchy Process:
AHP is a structured technique for dealing with complex decisions based on
mathematics and psychology. AHP is developed by Thomas L. Saaty in the
1970s and has been extensively studied and refined since then.
AHP develops priorities for different alternatives and according to criteria
judges the alternatives. Initially, priorities are sets according to importance
to achieve the goal, after that priority are derived for the performance of the
alternatives on each criteria, these priorities are derived based on pair-wise
assessments using judgments, or rations of measurements from a scales if
one exists. To make a decision there are three steps in AHP.
Step 1- Develop the weight for the criteria by developing a single pair-wise
comparison matrix for the criteria.
Step 2- Develop the rating for each decision alternative for each criteria.
Step 3- Calculate the weighted average rating for each decision alternatives,
choose the one with the highest score.
As show in figure.
10. PROBLEM IDENTIFICATION:
In Cloud computing environment resources may join or leave the cloud at
any instance of time. Therefore it is hard to realize the load balance among
the nodes offering processing power. Due to the dynamicity nature of
Cloud some resources may be overloaded or some may be underloaded.
Overloaded node decreases the performance so a better load scheduling is
necessary to achieve high performance in cloud computing environment or
we can say that load distribution is a critical factor to achieve high
performance in cloud computing environment load distribution may done
on the basis of some parameters like network parameters, application
characteristics and computing node capacity. If we consider single
parameter at a time then it may limit the overall performance. So we use
multiple parameter at a time.
When we use single parameter at a time it may limit the overall
performance of grid environment are listed below.
1. Network latency
After a limit performance increment rate consistently decreases
11. 2. Processor assignment
Less performance at higher no of applications
3. Transmission capacity
Better up to 100000 transactions and after that it become less effective
4. Inter process communication delay
Complexity increases as the no. of nodes in grid environment increases
5. Execution time
Performance is based on the clusters of similar and different types of jobs
6. CPU utilization
Variation in performance is large according to no. of jobs
12. PROPOSED OBJECTIVE
The objective of the model “Intelligent Multi Criteria Model for Load
Balancing in Cloud Environment” is to minimize the decision by using an
automatic decision making tool, analytical hierarchy process (AHP) and to
maximize the performance of cloud computing system by considering three
parameters simultaneously for load balancing decision.
There are three steps of our model:
Step 1- Cloud environment generation by using CloudSim Toolkit, a cloud
simulation toolkit based on java.
Step 2- Fetching the value of parameters named as available bandwidth,
processing speed of the node and number of parallel elements in the node.
Step 3- Automatic decision making by for load balancing using Analytical
hierarchy Process. AHP has a three step process to make a better decision.
13. :
Our Proposed model “Intelligent Multi Criteria Model for Load Balancing in
Cloud Environment” is explained in Figure
PROPOSED ALGORITHM/MODEL
14. ALGORITHM FOR MODEL GENERATION:
1. Cloud generation
{
CloudSim3.0.3
}
2. Parameter passing
{
Available Bandwidth
Number of Parallel Elements
Execution Time
}
3. Decision making for load balancing (using AHP)
{
Develop the weight for the criteria by developing a single pair-wise comparison matrix for
the criteria.
Develop the rating for each decision alternative for each criterion.
Calculate the weighted average rating for each decision alternatives; choose the one with the
highest score.
}
15. PUBLICATIONS:
Amandeep, Vandana, Faz “Different Strategies for Load Balancing in
Cloud Computing Environment: a critical Study” communicated to
International Journal of Scientific research & Technology, April
Edition.
REFERENCES:
[1] Sundararajan V.,: Scientific and Engineering Computing Group, Centre for
development of Advance Computing, Pune 411007, 2011.
[2] Nurmi, D., Wolski, R., Grzegorczyk, C., Obertelli, G., Soman, S., Youseff,
L., Zagorodnov, D.: The Eucalyptus Open-source Cloud-computing
System. CCGrid09: the 9th IEEE International Symposium on Cluster
Computing and the Grid, Shanghai, China (2009)
[3] Jaliya Ekanayake and Geoffrey Fox, High Performance Parallel
Computing with Clouds and Cloud Technologies
[4] A rmbrust, M., et al. Above the clouds: A Berkeley view of cloud
computing. Tech. Rep. UCB/EECS-2009-28, EECS Department, U.C.
Berkeley, Feb 2009.
[5] Rajkumar Buyyaa and et.al. Cloud computing and emerging IT platforms:
Vision, hype, and reality for delivering computing as the 5th utility
,ELSEVIER
16. [6] NIST: Nist definition of cloud computing
[7] Fox et al: Above the Clouds: A Berkeley View of Cloud computing feb 2009
[8] Nidhi Jain Kansal, Inderveer Chana, Cloud load balancing techniques-A Step
Towards Green Computing,IJCSI,vol 9,issue 1,Nov 2012
[9] Rajwinder Kaur and Pawan Luthra Load Balancing in Cloud Computing. Proc. of
Int. Conf. on Recent Trends in Information, Telecommunication and Computing,
ITC
[10] Ashish Revar, Malay Andhariya and Dharmendra Sutariya, Load Balancing in
Grid Environment using Machine Learning - Innovative Approach. International
Journal of Computer Applications (0975 – 8887) Volume 8– No.10, October 2010
[11] James Jasmine, Verma Bhupendra,: Efficient VM load balancing algorithm for a
cloud computing environment, Vol.4 No. 09 Sep 2012.
[12] A. Rajguru Abhijit and Apte S. S.: A comparative performance analysis of load
balancing algorithm in distributed system using qualitative parameters, IJRTE, Vol.-
1, Issue-3, August 2012.
[13] Galloway M. Jeffrey, Smith L. karl and Vrbsky S. Susan,: Power aware load
balancing for cloud computing, Proceeding of the world congress on engineering
and computer science 2011 vol. 1, Oct-2011.
[14] Sidhu Amandeepk Kaur, Kinger Supriya: Analysis of load balancing techniques in
cloud computing, IJCT, Vol.4, No.2, April-2013.
17. [15] The Analytic Hierarchy Process: www.dii.unisi.it/~mocenni/Note_AHP.pdf
[16] http://en.wikipedia.org/wiki/Analytic_hierarchy_process
[17] THE ANALYTIC HIERARCHY PROCESS, Geoff Coyle: Practical Strategy.
Open Access Material. AHP
[18] Tushar Desai and Jignesh Prajapati A Survey Of Various Load Balancing
Techniques And Challenges In Cloud Computing. INTERNATIONAL JOURNAL
OF SCIENTIFIC & TECHNOLOGY RESEARCH VOLUME 2, ISSUE 11,
NOVEMBER 2013
[19] R.R. Kotkondawar, P.A. Khaire, M.C. Akewar and Y.N. Patil, A Study of
Effective Load Balancing Approaches in Cloud Computing. International Journal of
Computer Applications , Volume 87 – No.8, February 2014
20] Rajesh George Rajan and V.Jeyakrishnan A Survey on Load Balancing in Cloud
Computing Environments. International Journal of Advanced Research in
Computer and Communication Engineering Vol. 2, Issue 12, December 2013
[21] Uddalak Chatterjee A Study on Efficient Load Balancing Algorithms in Cloud
Computing Environment. International Journal of Current Engineering and
Technology, 2013
[22] P.Warstein, H.Situ and Z.Huang(2010), “Load balancing in a cluster computer”
In proceeding of the seventh International Conference on Parallel and Distributed
Computing, Applications and Technologies, IEEE.