A New Approach for Dynamic Load Balancing Using Simulation In Grid Computing
0 likes55 views
This document proposes a new dynamic load balancing approach for grid computing using simulation. It discusses how dynamic load balancing algorithms can improve performance by reallocating tasks from heavily loaded nodes to lightly loaded nodes. The proposed approach implements a dynamic load balancing algorithm in a simulated grid environment. The algorithm uses information about current resource loads to schedule tasks in a way that aims to optimize resource usage and achieve high performance computing across the distributed grid resources.
![International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395 -0056
Volume: 03 Issue: 01 | Jan-2016 www.irjet.net p-ISSN: 2395-0072
_______________________________________________________________________________________
© 2016, IRJET ISO 9001:2008 Certified Journal Page 256
A NEW APPROACH FOR DYNAMIC LOAD BALANCING USING
SIMULATION IN GRID COMPUTING
Mr. Kapil B. Morey1, Prof. Sachin B. Jadhav2
1 PG Scholar, Computer Engineering Department, Padm. Dr. V.B. Kolte College of Engineering, Malkapur ,
Maharashtra, India
2 Assistant Professor, Computer Engineering Department, Padm. Dr. V.B. Kolte College of Engineering, Malkapur ,
Maharashtra, India
---------------------------------------------------------------------***---------------------------------------------------------------------
Abstract - Grid computing technology can be seen as
a positive alternative for implementing high-
performance distributed computing. The goal of Grid
computing is to create the illusion of virtual computer
out of a large collection of connected heterogeneous
nodes sharing various resources. The Grid system
needs competent load balancing algorithms for the
distribution of tasks in order to increase performance
and efficiency. The process of identifying requirements,
matching resources to applications, allocating those
resources, and scheduling and monitoring grid
resources over time in order to run grid applications
efficiently is known as Grid Resource Management. The
first phase of resource management is resource
discovery. The next step is monitoring and scheduling.
Monitoring keeps the details of the resources and
scheduling guides the job to appropriate resource. The
resources which are heavily loaded act as server of task
and the resources which are lightly loaded act as
receiver of task. Task relocation takes place from the
node which has high load towards the node which has
fewer loads. The main aim of load balancing is to
provide a distributed, low cost scheme that balances the
load across all the processors. In this paper a dynamic
load balancing algorithm is proposed in a simulated
grid environment which fulfils the objective to achieve
high performance computing by optimal usage of
geographically distributed and heterogeneous
resources in the grid environment.
Key Words: Grid, Resources, Simulation, Scheduling.
1. INTRODUCTION
The improvement of performance of computers and their
cost reduction is due to development in computing
resources. The recent researches on computing
architectures has resulted the emergence of a new
computing paradigm known as Grid computing. Grid is a
kind of distributed system which maintains the sharing
and synchronized use of geographically scattered and
multi owner resources autonomously from their physical
type and location, in dynamic implicit organizations that
share the same objective of solving large-scale
applications. The main aim is to prevent the condition
where some processors are overloaded with a set of tasks
while others are lightly loaded or even idle [4, 5].
The load balancing of the jobs in a grid environment can
considerably influence grid’s performance as the
resources are dynamic in nature. Hence the load of
resources changes with variation in configuration of grid.
A key characteristic of Grids is that resources like CPU
cycles and network capacities are shared among various
applications, and therefore, the amount of resources
available to any given application highly fluctuates over
time. Load balancing uses parallelism for boosting
throughput in order to minimize the response time by
intelligently distributing application resources. Load
balancing can be implemented using two techniques i.e.
static load balancing and dynamic load balancing. Static
load balancing algorithms assign the jobs 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. Dynamic load
balancing algorithms alter distribution of work during
runtime. Dynamic load balancing algorithms use the
updated current information of the load for taking
decisions about load distribution.
2. STRUCTURE OF A SIMPLE GRID
The simple grid can be defined as an interconnected
system for a distribution of non interactive workloads that
involve a large number of files. Grid can be built in all sizes
ranging from just a few machines to the group of machines
organized as a hierarchy spanning the world.](https://image.slidesharecdn.com/irjet-v3i144-171019085618/85/A-New-Approach-for-Dynamic-Load-Balancing-Using-Simulation-In-Grid-Computing-1-320.jpg)
![International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395 -0056
Volume: 03 Issue: 01 | Jan-2016 www.irjet.net p-ISSN: 2395-0072
_______________________________________________________________________________________
© 2016, IRJET ISO 9001:2008 Certified Journal Page 257
Fig -1: A simple Grid
The Simplest Grid consist of just a few machines all the
same hardware architecture and same operating system
connected on a local network. Because the machines have
the same architecture choosing application software for
these machines is usually simple. This type of structure
the grid is known as Intergrid which is homogeneous
systems. When the machines are been include to the
heterogeneous systems, more resources are available. File
sharing may still be accomplished using network file
systems. Such a grid is referred as an Intragrid [3].
3. DYNAMIC LOAD BALANCING
Load balancing algorithms can be defined by their
implementation of few policies [5]. The Information policy
states the workload of task information to be collected,
when it is to be collected and from where it is to be
collected. The next policy, Triggering policy determines
the appropriate period to start a load balancing operation.
The Resource type policy orders a resource as server or
receiver of tasks according to its availability status. The
Location policy uses the results of the resource type policy
to find a suitable partner for a server or receiver. The final
policy i.e. Selection policy determines the tasks to be
relocated from the node which has highly loaded
resources towards the node which has fewer loaded
resources. The load balancing algorithms are implemented
to allocate the tasks of a parallel program to workstations.
Multicomputer with dynamic load balancing allocates or
reallocates the resources at runtime based on task
information. In our approach Dynamic Load Balancing
Algorithm is implemented to multicomputer based on
resource type policy [5]. Load balancing feature can prove
invaluable for handling occasional peak loads of activity in
parts of a vast organization. There are important issues in
Load Balancing [4, 5]. If the Grid is already fully utilized,
the lowest priority work being performed on the Grid can
be temporarily suspended or even cancelled and
performed again later to make room for the higher
priority work.
The load arrangement in grid environment is altered by
some actions. These actions can be categorized as follows:
•Arrival of any new job and queuing of that job to any
particular node.
•Allocation of job to meticulous processor.
•Reallocating the job if load is unbalanced.
•Allocate the job to processor when it is free.
•Releasing the processor after it completes the entire job.
4. SCHEDULING AND LOAD BALANCING USING
GRID SIMULATION
4.1 Scheduling
The scheduling algorithms do not adequately address
congestion, and they do not take fairness considerations
into account. Fairness is needed for proper scheduling of
jobs. In Fair Scheduling, the jobs are allocated to multiple
processors so that the tasks with unsatisfied demand get
equal shares of time. The completion time of the jobs is
used to determine scheduling queue of the jobs. The
evaluation of completion time of the job is done by task
rate using a max min fair sharing algorithm. The job is
alloted to processor in accordance with growing degree of
completion time.
In scheduling algorithm, higher order tasks are completed
first which means that tasks are taken a higher priority
than the others which leads to starvation that increases
the completion time of tasks and load balance is not
guaranteed. To overcome this we put forward a load
balancing algorithm to provide unvarying load to the
resources so that all jobs are uniformly assigned to
processor depending on balanced fair rates. The main aim
of this algorithm is to reduce the overall time required to
complete the processing.
4.2 Dynamic Load Balancing Algorithm code
segment
Input: A set of R task and N number of processor with
computational capacity Pj
Output: A Schedule of R tasks
1. Construct a set of queues
2. q_size < R/N
3. While tasks present in queue do,
4. Assign demand rate of the task Xi
5. k=P/R
6. If Di < v
7. Assign Di to ith task as fair rate
8. Else](https://image.slidesharecdn.com/irjet-v3i144-171019085618/85/A-New-Approach-for-Dynamic-Load-Balancing-Using-Simulation-In-Grid-Computing-2-320.jpg)
![International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395 -0056
Volume: 03 Issue: 01 | Jan-2016 www.irjet.net p-ISSN: 2395-0072
_______________________________________________________________________________________
© 2016, IRJET ISO 9001:2008 Certified Journal Page 258
9. Assign v to ith task as fair rate
10. Calculate fair completion time t(D)
11. End while
12. End loop
13. Arrange the task in increasing order based on
their t(D) and submitted to processor
14. Calculate mean waiting time each scheduled task
15. If Zxy >0
16. Every processor having small amount capacity is
selected for relocation.
17. End If
18. End While.
4.3 Grid Simulation
In order to evaluate the performance of Grid resource
management and application scheduling algorithms, we
need to conduct repeatable and controlled experiments,
which are difficult due to Grid's inherent heterogeneity
and its dynamic nature. This indicates that all users need
to submit their tasks to the central scheduler that can be
targeted to perform comprehensive optimization such as
higher system utilization and overall satisfaction of user
depending on resource allocation policy or optimize for
high priority users.
Some of the GridSim features are outlined below:
•It allows modeling of different resource characteristics
and their failure properties; It enables simulation of
workload traces taken from real supercomputers;
•It supports reservation-based or auction mechanisms for
resource allocation; It allocates incoming jobs based on
space-or time-shared mode;
•It has the ability to schedule compute- and/or data-
intensive jobs;
•It has a background network traffic functionality based
on a probabilistic distribution.
6. Conclusion
This algorithm is able to provide sound results in terms of
make span and execution cost. Thus the algorithm assigns
the job to the available processors so that all requesting
jobs get identical time span which satisfies their demand.
We have described several aspects of load balancing
algorithm and established abundant notions which
illustrate its extensive potential through this projected
algorithm. This projected algorithm is definitely a capable
tendency to solve all types of load influenced problems
and high demanding applications. Objective of the grid
environment is to achieve high performance computing by
optimal usage of geographically distributed and But grid
application performance remains a challenge in dynamic
grid environment. At any moment resources can be
inhibited from grid and can be submitted to the grid.
ACKNOWLEDGEMENT
I am highly grateful to Prof. Sachin B. Jadhav, for his
sincere advice, encouragement and continuous guidance
in my work. I warmly acknowledge and express my special
thanks for him inspiring discussions and infallible
suggestions. I am also thankful to my family and
colleagues for their incessant support.
REFERENCES
[1] [Goswami& Das, 2015] Goswami S, Das A, “Deadline
stringency based job scheduling in computational
grid environment”, Computing for Sustainable Global
Development (INDIACom), 2015 2nd International
Conference, ISBN: 978-9-3805-4415-1, pp- 531-536,
IEEE March 2015.
[2] [Kinhekar& Gupta, 2014] Abhishek M. Kinhekar, Prof.
Hitesh Gupta, “A Review of Load Balancing in Grid
Computing”, International Journal of Advance
Research in Computer Science and Management
Studies, ISSN:2321-7782Vol. 2, Issue 8, pp-102-108,
August 2014.
[3] [Goswami& De Sarkar, 2013] Goswami S, De Sarkar
A, “A Comparative Study of Load Balancing
Algorithms in Computational Grid Environment”,
Computational Intelligence, Modelling and Simulation
(CIMSim), 2013 Fifth International Conference, ISBN:
978-1-4799-2308-3, pp- 99-104, IEEE September
2013.
[4] [Kaur& Singh, 2012] PawandeepKaur, Harshpreet
Singh, “Adaptive dynamic load balancing in grid
computing an approach,” International journal of
engineering science & advanced technology, ISSN:
2250–3676 Volume-2, Issue-3, 625 – 632, May-Jun
2012.
[5] [Srivastava, 2011] Prabhat Srivastava, “Improving
Performance in Load Balancing Problem on the Grid
Computing System”, International Journal of
Computer Applications (0975 – 8887), Volume 16–
No.1, February 2011.
BIOGRAPHIES
Mr. Kapil B. Morey
PG Scholar, Computer Engineering Dept.,
Padm. Dr. V.B. Kolte College of Engg.,
Malkapur, Maharashtra, India.](https://image.slidesharecdn.com/irjet-v3i144-171019085618/85/A-New-Approach-for-Dynamic-Load-Balancing-Using-Simulation-In-Grid-Computing-3-320.jpg)
A New Approach for Dynamic Load Balancing Using Simulation In Grid Computing
- 1. International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395 -0056 Volume: 03 Issue: 01 | Jan-2016 www.irjet.net p-ISSN: 2395-0072 _______________________________________________________________________________________ © 2016, IRJET ISO 9001:2008 Certified Journal Page 256 A NEW APPROACH FOR DYNAMIC LOAD BALANCING USING SIMULATION IN GRID COMPUTING Mr. Kapil B. Morey1, Prof. Sachin B. Jadhav2 1 PG Scholar, Computer Engineering Department, Padm. Dr. V.B. Kolte College of Engineering, Malkapur , Maharashtra, India 2 Assistant Professor, Computer Engineering Department, Padm. Dr. V.B. Kolte College of Engineering, Malkapur , Maharashtra, India ---------------------------------------------------------------------***--------------------------------------------------------------------- Abstract - Grid computing technology can be seen as a positive alternative for implementing high- performance distributed computing. The goal of Grid computing is to create the illusion of virtual computer out of a large collection of connected heterogeneous nodes sharing various resources. The Grid system needs competent load balancing algorithms for the distribution of tasks in order to increase performance and efficiency. The process of identifying requirements, matching resources to applications, allocating those resources, and scheduling and monitoring grid resources over time in order to run grid applications efficiently is known as Grid Resource Management. The first phase of resource management is resource discovery. The next step is monitoring and scheduling. Monitoring keeps the details of the resources and scheduling guides the job to appropriate resource. The resources which are heavily loaded act as server of task and the resources which are lightly loaded act as receiver of task. Task relocation takes place from the node which has high load towards the node which has fewer loads. The main aim of load balancing is to provide a distributed, low cost scheme that balances the load across all the processors. In this paper a dynamic load balancing algorithm is proposed in a simulated grid environment which fulfils the objective to achieve high performance computing by optimal usage of geographically distributed and heterogeneous resources in the grid environment. Key Words: Grid, Resources, Simulation, Scheduling. 1. INTRODUCTION The improvement of performance of computers and their cost reduction is due to development in computing resources. The recent researches on computing architectures has resulted the emergence of a new computing paradigm known as Grid computing. Grid is a kind of distributed system which maintains the sharing and synchronized use of geographically scattered and multi owner resources autonomously from their physical type and location, in dynamic implicit organizations that share the same objective of solving large-scale applications. The main aim is to prevent the condition where some processors are overloaded with a set of tasks while others are lightly loaded or even idle [4, 5]. The load balancing of the jobs in a grid environment can considerably influence grid’s performance as the resources are dynamic in nature. Hence the load of resources changes with variation in configuration of grid. A key characteristic of Grids is that resources like CPU cycles and network capacities are shared among various applications, and therefore, the amount of resources available to any given application highly fluctuates over time. Load balancing uses parallelism for boosting throughput in order to minimize the response time by intelligently distributing application resources. Load balancing can be implemented using two techniques i.e. static load balancing and dynamic load balancing. Static load balancing algorithms assign the jobs 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. Dynamic load balancing algorithms alter distribution of work during runtime. Dynamic load balancing algorithms use the updated current information of the load for taking decisions about load distribution. 2. STRUCTURE OF A SIMPLE GRID The simple grid can be defined as an interconnected system for a distribution of non interactive workloads that involve a large number of files. Grid can be built in all sizes ranging from just a few machines to the group of machines organized as a hierarchy spanning the world.
- 2. International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395 -0056 Volume: 03 Issue: 01 | Jan-2016 www.irjet.net p-ISSN: 2395-0072 _______________________________________________________________________________________ © 2016, IRJET ISO 9001:2008 Certified Journal Page 257 Fig -1: A simple Grid The Simplest Grid consist of just a few machines all the same hardware architecture and same operating system connected on a local network. Because the machines have the same architecture choosing application software for these machines is usually simple. This type of structure the grid is known as Intergrid which is homogeneous systems. When the machines are been include to the heterogeneous systems, more resources are available. File sharing may still be accomplished using network file systems. Such a grid is referred as an Intragrid [3]. 3. DYNAMIC LOAD BALANCING Load balancing algorithms can be defined by their implementation of few policies [5]. The Information policy states the workload of task information to be collected, when it is to be collected and from where it is to be collected. The next policy, Triggering policy determines the appropriate period to start a load balancing operation. The Resource type policy orders a resource as server or receiver of tasks according to its availability status. The Location policy uses the results of the resource type policy to find a suitable partner for a server or receiver. The final policy i.e. Selection policy determines the tasks to be relocated from the node which has highly loaded resources towards the node which has fewer loaded resources. The load balancing algorithms are implemented to allocate the tasks of a parallel program to workstations. Multicomputer with dynamic load balancing allocates or reallocates the resources at runtime based on task information. In our approach Dynamic Load Balancing Algorithm is implemented to multicomputer based on resource type policy [5]. Load balancing feature can prove invaluable for handling occasional peak loads of activity in parts of a vast organization. There are important issues in Load Balancing [4, 5]. If the Grid is already fully utilized, the lowest priority work being performed on the Grid can be temporarily suspended or even cancelled and performed again later to make room for the higher priority work. The load arrangement in grid environment is altered by some actions. These actions can be categorized as follows: •Arrival of any new job and queuing of that job to any particular node. •Allocation of job to meticulous processor. •Reallocating the job if load is unbalanced. •Allocate the job to processor when it is free. •Releasing the processor after it completes the entire job. 4. SCHEDULING AND LOAD BALANCING USING GRID SIMULATION 4.1 Scheduling The scheduling algorithms do not adequately address congestion, and they do not take fairness considerations into account. Fairness is needed for proper scheduling of jobs. In Fair Scheduling, the jobs are allocated to multiple processors so that the tasks with unsatisfied demand get equal shares of time. The completion time of the jobs is used to determine scheduling queue of the jobs. The evaluation of completion time of the job is done by task rate using a max min fair sharing algorithm. The job is alloted to processor in accordance with growing degree of completion time. In scheduling algorithm, higher order tasks are completed first which means that tasks are taken a higher priority than the others which leads to starvation that increases the completion time of tasks and load balance is not guaranteed. To overcome this we put forward a load balancing algorithm to provide unvarying load to the resources so that all jobs are uniformly assigned to processor depending on balanced fair rates. The main aim of this algorithm is to reduce the overall time required to complete the processing. 4.2 Dynamic Load Balancing Algorithm code segment Input: A set of R task and N number of processor with computational capacity Pj Output: A Schedule of R tasks 1. Construct a set of queues 2. q_size < R/N 3. While tasks present in queue do, 4. Assign demand rate of the task Xi 5. k=P/R 6. If Di < v 7. Assign Di to ith task as fair rate 8. Else
- 3. International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395 -0056 Volume: 03 Issue: 01 | Jan-2016 www.irjet.net p-ISSN: 2395-0072 _______________________________________________________________________________________ © 2016, IRJET ISO 9001:2008 Certified Journal Page 258 9. Assign v to ith task as fair rate 10. Calculate fair completion time t(D) 11. End while 12. End loop 13. Arrange the task in increasing order based on their t(D) and submitted to processor 14. Calculate mean waiting time each scheduled task 15. If Zxy >0 16. Every processor having small amount capacity is selected for relocation. 17. End If 18. End While. 4.3 Grid Simulation In order to evaluate the performance of Grid resource management and application scheduling algorithms, we need to conduct repeatable and controlled experiments, which are difficult due to Grid's inherent heterogeneity and its dynamic nature. This indicates that all users need to submit their tasks to the central scheduler that can be targeted to perform comprehensive optimization such as higher system utilization and overall satisfaction of user depending on resource allocation policy or optimize for high priority users. Some of the GridSim features are outlined below: •It allows modeling of different resource characteristics and their failure properties; It enables simulation of workload traces taken from real supercomputers; •It supports reservation-based or auction mechanisms for resource allocation; It allocates incoming jobs based on space-or time-shared mode; •It has the ability to schedule compute- and/or data- intensive jobs; •It has a background network traffic functionality based on a probabilistic distribution. 6. Conclusion This algorithm is able to provide sound results in terms of make span and execution cost. Thus the algorithm assigns the job to the available processors so that all requesting jobs get identical time span which satisfies their demand. We have described several aspects of load balancing algorithm and established abundant notions which illustrate its extensive potential through this projected algorithm. This projected algorithm is definitely a capable tendency to solve all types of load influenced problems and high demanding applications. Objective of the grid environment is to achieve high performance computing by optimal usage of geographically distributed and But grid application performance remains a challenge in dynamic grid environment. At any moment resources can be inhibited from grid and can be submitted to the grid. ACKNOWLEDGEMENT I am highly grateful to Prof. Sachin B. Jadhav, for his sincere advice, encouragement and continuous guidance in my work. I warmly acknowledge and express my special thanks for him inspiring discussions and infallible suggestions. I am also thankful to my family and colleagues for their incessant support. REFERENCES [1] [Goswami& Das, 2015] Goswami S, Das A, “Deadline stringency based job scheduling in computational grid environment”, Computing for Sustainable Global Development (INDIACom), 2015 2nd International Conference, ISBN: 978-9-3805-4415-1, pp- 531-536, IEEE March 2015. [2] [Kinhekar& Gupta, 2014] Abhishek M. Kinhekar, Prof. Hitesh Gupta, “A Review of Load Balancing in Grid Computing”, International Journal of Advance Research in Computer Science and Management Studies, ISSN:2321-7782Vol. 2, Issue 8, pp-102-108, August 2014. [3] [Goswami& De Sarkar, 2013] Goswami S, De Sarkar A, “A Comparative Study of Load Balancing Algorithms in Computational Grid Environment”, Computational Intelligence, Modelling and Simulation (CIMSim), 2013 Fifth International Conference, ISBN: 978-1-4799-2308-3, pp- 99-104, IEEE September 2013. [4] [Kaur& Singh, 2012] PawandeepKaur, Harshpreet Singh, “Adaptive dynamic load balancing in grid computing an approach,” International journal of engineering science & advanced technology, ISSN: 2250–3676 Volume-2, Issue-3, 625 – 632, May-Jun 2012. [5] [Srivastava, 2011] Prabhat Srivastava, “Improving Performance in Load Balancing Problem on the Grid Computing System”, International Journal of Computer Applications (0975 – 8887), Volume 16– No.1, February 2011. BIOGRAPHIES Mr. Kapil B. Morey PG Scholar, Computer Engineering Dept., Padm. Dr. V.B. Kolte College of Engg., Malkapur, Maharashtra, India.