Cloud Partitioning of Load Balancing Using Round Robin Model

ISSN (Online): 2349-7084
GLOBAL IMPACT FACTOR 0.238

ISRA JIF 0.351
INTERNATIONAL JOURNAL OF COMPUTER ENGINEERING IN RESEARCH TRENDS
VOLUME 1, ISSUE 6, DECEMBER 2014, PP 367-372
Cloud Partitioning of Load Balancing Using

Round Robin Model
1
M.V.L.SOWJANYA, 2 D.RAVIKIRAN
1
M.Tech Research Scholar, Priyadarshini Institute of Technology and Science for Women
2
Professor, Priyadarshini Institute of Technology and Science for Women
Abstract: The purpose of load balancing is to look up the performance of a cloud environment through an appropriate
circulation strategy. Good load balancing will construct cloud computing for more stability and efficiency. This paper
introduces a better round robin model for the public cloud based on the cloud partitioning concept with a switch mechanism
to choose different strategies for different situations. Load balancing is the process of giving out of workload among
different nodes or processor. It will introduces a enhanced approach for public cloud load distribution using screening and
game theory concept to increase the presentation of the system.
Key words: load balancing model; public cloud; cloud partition; game theory
1. INTRODUCTION heavily dependent upon dividing up work effectively

In the cloud technology users will introduce new types across the several participating nodes.
of services in that we have to change our life. Without A round robin is an arrangement of choosing all
paying an awareness of details we get new services [2]. elements in a group equally in some rational order,
NIST gave a definition of cloud computing as a model usually from the top to the bottom of a list and then
for enabling everywhere, expedient, on-demand starting again at the top of the list and so on. A simple
network access to a collective pool of configurable way to think of round robin is that it is about "taking
computing resources (e.g., networks, servers, storage, turns. Thus, this model divides the public cloud into
applications, and services) that can be swiftly several cloud partitions. Load balancing simplification
delivered and released with minimal management is based on environment is very large and complex.
endeavor or service provider interaction [3]. More and The cloud has a main controller that chooses the
more people pay interest to cloud computing [4, 5]. suitable partitions for arriving jobs while the balancer
Cloud computing is systematic and scalable but for each cloud partition chooses the best load balancing
preserving the stability of processing so many jobs in strategy.
the cloud computing environment is a very complex
problem with load balancing receiving much attention 2. RELATED WORK
for researchers. We get more studies of load balancing for the cloud
environment. Alder was explained the Load balancing
In a job arrival pattern also not expected and the in cloud computing who began the tools and methods
capability of each node cloud be different, for load normally used for load them balancing in the cloud. In
balancing problem, workload control is very important Chaczko et al [8] report arrives a new problem in load
to improve system performance and sustain stability. balancing in the cloud that needs new architectures to
Load balancing is the development of attractive adapt to many modifications. In that they were
performance of a parallel and distributed system via explained about load balancing plays in improving the
the redistribution of load among different processing presentation and maintaining strength.
units or nodes. Load balancing is corresponding to, In Here we are discussing about load balancing
the scenario of circulated network environment of algorithms, such as Round Robin, Equally Spread
computing hosts, the functioning of the system is Current Execution Algorithm, and Throttled Load
IJCERT2014 367
www.ijcert.org
ISRA JIF 0.351
Balancing. Balancer has multiple servers attached to it. It keeps

Nishant et al [9] used the Throttled load optimization the record of all status information. Initially a request
method in nodes load balancing. Randles et al.[10] arrives at the system. The best partition search strategy
checking the performance time and cost. He gave a
helps to decide to which partition the request has to be
compared analysis of some algorithms in cloud
assigned. The whole process is shown in Fig.2.
computing. In this scenario concluded that the ESCE
algorithm and throttled algorithm are better than the
Round Robin algorithm. Some of the classical loads 3.1 Main controller and balancers
balancing methods are similar to the allowance
The main controller and the balancers have done the
method in the operating system, for example, the
load balancing solution. Main controller distributes the
Round Robin algorithm and the First Come First
planned jobs to the suitable cloud partition based on
Served (FCFS) rules. The Round Robin algorithm is
execution time of the job. Main controller contains the
used here because it is fairly simple.
configuration information of each partition. Based on
the job length the main controller decides to which
3. SYSTEM MODEL partition he job has to be assigned to get the response
Cloud architectures are mainly divided into three in less time. The balancers in each partition collect the
categories: public, private, and partner. The most status information from every node and then choose
familiar model of cloud computing to many consumers the right plan to distribute the jobs. In Fig.3 we can see
a relationship between the balancers and the main
is the public cloud model, under which cloud services
controller.
are provided in a virtualized environment, constructed
using pooled shared physical resources, and accessible 3.2 Assigning jobs to the cloud partition
over a public network such as the internet. Public
The first step is to choose the right partition after the
clouds, however, provide services to multiple clients
job arrives at the public cloud. The cloud partition
using the same shared infrastructure. Public cloud has
status can be divided into three types:
contains subarea that is cloud computing with
partitions based on the geographic locations. The
architecture is shown in Fig.1.
Cloud partitioning model follow the load balancing
strategy. Load balancing will start after creation of
cloud partitioning: The cloud has a main controller
(MC) which decides the suitable partitions for new
jobs.
Fig. 2 Job assignment strategy.
(1) Overload: When the percentage of the

Fig.1 typical cloud partitioning. overloaded nodes exceeds gamma, change to
overloaded status.
IJCERT2014 368
www.ijcert.org
ISRA JIF 0.351
(2) Normal: When the percentage of the normal are related to node load degree. The static parameters
nodes exceeds beta, change to normal load contain the number of CPUs, the CPU giving out
status. speeds, the memory size, etc. Dynamic parameters are
(3) Idle: When the percentage of idle nodes
the memory use ratio, the CPU use ratio, the network
exceeds alpha, change to idle status.
bandwidth, etc. The load degree is calculated from
The status information will refreshed by the main these parameters as below:
controller. It has to communicate with the balancers
frequently. The main controller then transmits the jobs Step 1 Define a load parameter set: Define a load
using the following approach: parameter set: F = ,F1; F2;<; Fm} with each Fi(1 i
The main controller inquiries the cloud partition where m; Fi *0, 1+) parameter being either static or
job is located when the job i appears at the system. If dynamic.
this locations status is idle or normal, the job is
Step 2 Compute the load degree as:
switched locally. If not, another cloud partition is
found that is not overloaded. The algorithm is shown Load degree (N) =i=1m i Fi
in Algorithm 1. i (i=1n i = 1) are weights that may differ for
different kinds of jobs. N represents the current
3.3 Assigning jobs to the nodes in the node.
Step 3 Define estimate benchmarks. Calculate the
cloud partition
average cloud partition degree from the node load
The cloud partition balancer collects load information degree statistics as:
from every node to evaluate the cloud partition status. Load_degreeavg = i=1n load_degree (Ni) / n
This evaluation of each nodes load status is very
important. The first task is to define the load degree of The bench mark Load_degreehigh is then set for
each nodes. different situations based on the Load_degreeavg.
Step 4 Three nodes load status levels are then defined
Algorithm 1 Best Partition Searching as:
Begin Idle When

Load_degree (N) = 0;
While job do In this node there is no job being processed so the
Search BestPartition status is charged to Idle.
(job); Normal For
0 < Load_degree (N) Load_degreehigh,
If partitionState == idle || partition State ==
the node is normal and it can process other jobs.
normal then Overloaded When
Load_degreehigh Load degree (N);
Send Job to Partition;
The node does not exist and not obtain jobs until it
Else returns to the normal.The cloud partition balancers are
Search for another created a load status tables it takes an input from load
degree results. Each every fixed period T we get
Partition; balancer it has a Load Status Table and refreshes. If
you want to calculate the partition status you can use
End if balancer of tables. Each partition position has a
End different load balancing solution. When a job appears
at a cloud partition, the balancer gives the job to the
while nodes based on its current load approach. This
approach is changed by the balancers as the cloud
End
partition position changes.
The various static parameters and dynamic parameters
IJCERT2014 369
www.ijcert.org
ISRA JIF 0.351
4. CLOUD PARTITION LOAD BALANCING walks throughout the queue again and again. Jobs will
then be allotted to nodes with low load degrees. The
STRATEGY node order will be altered when the balancer recharges
4.1 Motivation the Load condition Table. Still, there may be examine
and mark difference at the refresh period T. When the
We can improve the performance of the entire cloud balance table is revived, at this immediate, if a job
based on good load balance. However, there is no appears at the cloud partition, it will bring the
common technique that can adapt to all possible contradictory problem. The system situation will have
different conditions. Various techniques have been changed but the information will still used. This may
developed in civilizing existing solutions to resolution lead to an incorrect load approach option and an
new problems. incorrect nodes order. To determine this problem, two
Each particular technique has advantage in exacting Load category Tables should be created as: Load Status
area but not in all conditions. So, the current model Table_1 and Load Status Table_2. A flag is also
combines several techniques and switches between the allocated to each table to specify examine or mark.
load balance techniques based on the system condition.
A relatively simple technique can be used for the When the flag = Read, then the Round Robin
partition idle state with a more difficult technique for based on the load degree evaluation algorithm is using
the normal state. The load balancers at that time switch this table.
methods as the status changes. Now, the idle status
uses an enhanced Round Robin algorithm whereas the When the flag = Write, the table is mortal
normal status applies a game theory based load refreshed, new information is written into this table.
balancing approach.
Thus, at each instant, one table offers the correct
4.2 Load balance strategy for the idle node locations in the queue for the enhanced Round
status Robin algorithm, while the other is mortal arranged
with the efficient information. Once the data is
Many computing sources are available and fairly few recharged, the table flag is changed to Read and the
tasks are incoming while the cloud partition is idle. In other tables flag is changed to Write. The two tables
this situation, this cloud partition has the capability to then alternate to solve the variation. The
development tasks as fast as possible so a simple load
balancing method can be used.
Here we can see many simple load balance algorithm
techniques such as the Random algorithm, the Weight
Round Robin, and the Dynamic Round Robin [12].
Here, the Round Robin algorithm maintains simplicity.
The Round Robin algorithm is one of the simplest

load balancing algorithms, which exceeds each new
demand to the next server in the queue. The algorithm
does not record the condition of every connection thus
it has no condition information. In the normal Round
Robin algorithm, every node has an equal chance to be
special. However, in a public cloud, the design and the
presentation of each node will be not the same; thus, Process is shown in Fig.3.
this method may overload some nodes. Thus, an
improved Round Robin algorithm is old, which called 4.3 Load balancing strategy for the
Round Robin based on the load degree evaluation.
normal status
The algorithm is still literally simple. Before the Round
Robin step, the nodes in the load balancing table are
The jobs are arriving much faster than in the idle state
planned based on the load degree from the lowest to
and the position is far more difficult, so a different
the highest. The system makes a circular queue and
IJCERT2014 370
www.ijcert.org
ISRA JIF 0.351
approach is used for the load balancing while the cloud each job. The plan then changes as the nodes position
partition is normal. Each user desires his tasks finished modify.
in the shortest time, so the public cloud requires
methods that can entire jobs of all users with logical 5. FUTURE WORKS
response time.
While this work is now a conceptual framework,
additional work is desired to apply the framework and
Penmatsa and chronopoulos planned a static load
determine new problems. Some significant points are:
balancing approach based on game theory for
distributed systems. And this work provides us with a (1) Cloud division rules: Cloud division is not a easy
new analysis of the load balance problem in the cloud problem. Thus, the framework will require a
atmosphere. As an implementation of distributed comprehensive cloud division methodology. For
system, the load balancing in the cloud computing example, nodes in a group may be extreme from
atmosphere can be viewed as a game. other nodes or there will be some groups in the
same geographic area that are still far apart. The
division rule should only base on the geographic
Game theory has non-supportive games and
location (region or state).
supportive games. In cooperative games, the result
makers ultimately come to an contract which is called a (2) Find other load balance strategy: Other load
binding agreement. Each decision maker decides by balance approaches may provide better results, so
comparing notes with each others. In non-cooperative tests are required to balance different approaches.
games, each decision maker makes results only for his Many tests are required to pledge system
own benefit. The system then makes the Nash availability and efficiency.
equilibrium, where each result maker makes the (3) A better load status evaluation: A good algorithm
optimized decision. The Nash equilibrium is when is desired to set Load_degree high and
each player in the game has selected approach and no Load_degree low, and the estimate device requires
player can advantage by changing his or her approach being more widespread.
while the other players approaches stay unchanged. (4) How to locate the refresh period: In the data
statistics analysis, the main controller and the
There have been many studies in using game theory
cloud partition balancers require to restore the
for the load balancing. Grosu et al.[14] designed a load
information at a fixed period. If the period is too
balancing approach based on game theory for the
short, the high frequency will manipulate the
distributed systems as a non-supportive game using
system presentation. If the period is too long, the
the distributed structure. They balanced this algorithm
information will be too previous to make good
with other conventional methods to show that their
conclusion. Thus, tests and statistical tools are
algorithm was less difficulty with better presentation.
desired to set logical refresh periods.
Aote and Kharat [15] gave a dynamic load balancing
model based on game theory. This model is linked on
the dynamic load position of the system with the users
6. CONCLUSSION
being the choice makers in a non-supportive game.
This paper introduces a better round robin model for
Since the grid computing and cloud computing the public cloud based on the cloud partitioning
atmospheres are also distributed system, these
concept with a switch mechanism to choose different
algorithms can also be used in grid computing and
strategies for different situations. Load balancing is the
cloud computing atmospheres. Earlier studies have
exposed that the load balancing approach for a cloud process of giving out of workload among different
partition in the normal load condition can be viewed as nodes or processor. It will introduces a enhanced
a non-supportive game, as explained here. approach for public cloud load distribution using
screening and game theory concept to increase the
In this model, the most important step is finding the presentation of the system. Load balancing will
suitable value of sji. The existing model uses the
construct cloud computing for more stability and
method of Grosu et al.*14+ called the best reply to
efficiency
calculate sji of each node, with a greedy algorithm then
used to analyze sji for all nodes. This method gives the
Nash equilibrium to minimize the response time of
IJCERT2014 371
www.ijcert.org
ISRA JIF 0.351
[13] S. Penmatsa and A. T. Chronopoulos, Game-

theoretic static balancing for distributed systems,
REFERENCES load journal of Distributed computing pp.502-11-
[1] R. Hunter, The why of cloud, 42.
http://www.gartner.com/DisplayDocument?doc [14] D. Grosu, A. T. Chronopoulos, and M. Y. Leung,
cd=226469&ref= gnoreg, 2012. Load balancing in distributed systems: An
approach using cooperative games, in Proc. 16th
[2] M. D. Dikaiakos, D. Katsaros, P. Mehra, G. Pallis,
IEEE Intl. Parallel and Distributed Processing
and A. Vakali, Cloud computing: Distributed
Symp., Florida, USA, Apr. 2002, pp. 52-61.
internet computing for IT and scientific research,
[15] S. Aote and M. U. Kharat, A game-theoretic
Internet Computing, vol.13, no.5, pp.10-13, Sept.-
model for dynamic load balancing in distributed
Oct. 2009.
systems, in Proc. the International Conference on
[3] P. Mell and T. Grance, The NIST definition of
Advances in Computing, Communication and
cloud computing, http://csrc.nist.gov/
Control (ICAC3 09), New York, USA, 2009, pp.
publications/nistpubs/800-145/SP800-145.pdf,
235-238.
2012.
[4] Microsoft Academic Research, Cloud computing,
http:// libra.msra.cn/Keyword/6051/cloud-
computing? query = cloud%20computing, 2012.
[5] Google Trends, Cloud computing,
http://www.google.com/trends/explore#q=cloud
%20computing, 2012.
[6] N. G. Shivaratri, P. Krueger, and M. Singhal, Load
distributing for locally distributed systems,
Computer, vol. 25, no. 12, pp. 33-44, Dec. 1992.
[7] B. Adler, Load balancing in the cloud: Tools, tips
and techniques, http://www.rightscale. com/info
center/white-papers/Load-Balancing-in-the-
Cloud.pdf, 2012
[8] Z. Chaczko, V. Mahadevan, S. Aslanzadeh, and C.
Mcdermid, Availability and load balancing in
cloud computing, presented at the 2011
International Conference on Computer and
Software Modeling, Singapore, 2011.
[9] K. Nishant, P. Sharma, V. Krishna, C. Gupta, K. P.

Singh, International Conference on Computer
Modeling and Simulation (UKSim), Cambridge
shire, United Kingdom, Mar. 2012, pp. 28-30.
[10] M. Randles, D. Lamb, and A. Taleb-Bendiab, A
comparative study into distributed load balancing
algorithms for cloud computing, in Proc. IEEE
24th International Conference on Advanced
Information Networking and Applications, Perth,
Australia, 2010, pp. 551-556.
[11] A. Rouse, Public cloud,
http://searchcloudcomputing.techtarget.com/defi
nition/public-cloud, 2012.
[12] D. MacVittie, Intro to load balancing for
developers the algorithms,
https://devcentral.f5.com/blogs/us/intro-to-load-
balancing-for-developers-ndash-the-algorithms,
2012.
IJCERT2014 372
www.ijcert.org

Cloud Partitioning of Load Balancing Using Round Robin Model

Uploaded by

Cloud Partitioning of Load Balancing Using Round Robin Model

Uploaded by

ISSN (Online): 2349-7084

GLOBAL IMPACT FACTOR 0.238

Cloud Partitioning of Load Balancing Using

1. INTRODUCTION heavily dependent upon dividing up work effectively

Balancing. Balancer has multiple servers attached to it. It keeps

Fig. 2 Job assignment strategy.

(1) Overload: When the percentage of the

Begin Idle When

The various static parameters and dynamic parameters

The Round Robin algorithm is one of the simplest

[13] S. Penmatsa and A. T. Chronopoulos, Game-

[9] K. Nishant, P. Sharma, V. Krishna, C. Gupta, K. P.

You might also like