Download as pdf or txt
Download as pdf or txt
You are on page 1of 7

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/329911814

Multi-Criteria Strategy for Job Scheduling and Resource Load Balancing in


Cloud Computing Environment

Conference Paper · May 2015

CITATIONS READS

0 47

5 authors, including:

Aminu Mahmoud Zarina Mohamad


Yusuf Maitama Sule University Universiti Sultan Zainal Abidin | UniSZA
5 PUBLICATIONS 18 CITATIONS 34 PUBLICATIONS 128 CITATIONS

SEE PROFILE SEE PROFILE

Fadhilah Ahmad Mohamad A Mohamed


Universiti Sultan Zainal Abidin | UniSZA Universiti Sultan Zainal Abidin
23 PUBLICATIONS 182 CITATIONS 152 PUBLICATIONS 1,204 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Aminu Mahmoud on 17 January 2019.

The user has requested enhancement of the downloaded file.


The 4th International Conference on Computer Science and Computational Mathematics (ICCSCM 2015)

Multi-Criteria Strategy for Job Scheduling and


Resource Load Balancing in Cloud Computing
Environment
Aminu Abdulkadir Mahmoud, Zarina Mohamad, Wan Nor Shuhadah Wan Nik and Fadhilah Ahmad
Department of Computer Science,
Faculty of Informatics and Computing,
Universiti Sultan Zainal Abidin, Malaysia

{si0453@putra., zarina@, wnshuhadah@, fad@}unisza.edu.my

Abstract: Cloud computing is growing rapidly over the years and it Cloud environment consists of huge amount of data,
faces challenges especially in resource management. Resource distributed storage and demand of good services; which can
management in cloud computing is necessary due to its distributed provide high reliability and availability fairness of services
nature with different user demands. Quality of Service (QoS), load and better access performance over the internet.
balancing and throughput are identified as some of the benefits of
proper resource management. This research focuses on job
Cloud computing resources need to be managed and utilized
scheduling and resource load balancing in cloud environment. We properly to achieve desired results. This research focused on
proposed an efficient algorithm based on multi-criteria strategy. job scheduling as part of cloud resource management. We
The algorithm consists of two main phases. In the first phase the consider the distributed nature of cloud environment where
shortest job completion time is measured based on the completion different datacenters may be located at different geographical
time of three techniques i.e. min-min, max-min and suffrage. locations with their hosts and virtual machines (VMs). In
Meanwhile in the second phase genetic algorithm is implemented addition to the characteristics of datacenters, hosts and VMs,
for resource load balancing. CloudSim simulator is used to measure an efficient algorithm needs to be proposed that can enhance
the performance and efficiency of the proposed algorithm. The job scheduling and resource load balancing. The proposed
proposed algorithm enhances jobs scheduling and resource load
balancing by ensuring an efficient utilization of the available
algorithm implements three scheduling techniques i.e. min-
resources. min, max-min and suffrage to find the minimum makespan.
Furthermore, the proposed algorithm will also take
Keywords: Cloud Computing, Genetic Algorithm, Job Scheduling, advantage on the benefits of genetic algorithm for resource
Virtual Machine, Load Balancing. load balancing. The idea on balancing the resource workload
will indirectly improve job response time, resource
availability, resource throughput and of course the
1. Introduction performance of the cloud environment as whole. This can be
achieved by avoiding extreme utilization (i.e. under
Cloud Computing is a new area of computing and interested utilization or over utilization) of some cloud resources.
field of research nowadays [1]. Cloud Computing is a term
referring to applications, development platforms,
infrastructures, storages, communications and collaborations
over the internet on-demand via services offered by cloud
2. Related Works
providers [2, 4 and 9]. However Cloud computing has There are several publications and research papers on this
become an important pillar in the field of IT. Cloud new field of modern day computing with a lot of
computing in modern days computing environment can expectations on its deliverance. This is especially true when
promise number of benefits by providing easy and affordable it aims to minimize the makespan of the job(s). Research in
cost of IT equipment in terms of software and hardware. [10] proposed hosts scheduling by implementing the
Furthermore, it can also reduce the cost of outsourcing, data Suffrage algorithm coupled with Genetic Algorithm. The
backup and scarce IT talent by providing steady access for objective of the research is to minimize the make span of the
cloud users [5]. job(s).
Despite of the benefits and expectations Cloud computing Another research in [11] was conducted with the aim of
can offer, it has some grey areas that need to be improved. addressing load balancing problems in VM resources
This is especially true when cloud computing environment is scheduling. In this research, job scheduling strategy is
heterogeneous in nature and its resources are geographically proposed in order to load balance the VM resources based on
distributed. genetic algorithm. In addition it also addressed the problems
Cloud computing brings together large amount of data which of high migration cost after jobs are scheduled on VMs.
needs to be efficiently managed in order to provide on Meanwhile research in [12], proposed an algorithm using
demand services. Job scheduling is a part of data two conventional task scheduling algorithm i.e. min-min and
management in cloud environment which is associated with max-min algorithms. These algorithms uses certain heuristics
numerous challenges due to the nature of cloud environment. to select between the two conventional algorithms so that the

176
The 4th International Conference on Computer Science and Computational Mathematics (ICCSCM 2015)

overall make span is minimized and the tasks are scheduled performance of the proposed strategy was compared with
on machines in either space shared or time shared manner First Come First Serve (FCFS) and Round Robin Scheduling
and it was compared with FCFS scheduling technique. techniques. When tested in CloudSim toolkit, the result
Furthermore, research in [13] based on job scheduling in shows that it gives better performance than other traditional
cloud computing environment, where a generalized priority scheduling algorithms.
algorithm was proposed for efficient job execution. The

Table 1: Summary of Scheduling Techniques in Cloud Computing


Author / Year Method and Issue Strength Weakness
Goyal T. and The researcher proposed the The proposed algorithm is The proposed algorithm does
Agrawal A., 2013, used of Suffrage coupled with able to find minimum not concern about load
[10] Genetic Algorithm. completion time for a balancing between the virtual
given task(s). machines.
The objective of the research is
to minimize the make span of
the job(s)
Gu J., Hu J., Zhao T. The genetic algorithm was This strategy is able to The strategy does not take into
and Sun G., 2012, proposed in order to address the solve the problem of load consideration the minimum
[11] problems of load balancing and imbalance and high completion time of a given
high migration cost. migration cost task(s).

The aim of this research is to


address the problems of load
balancing in Virtual Machine
(VM) resource scheduling.
Katyal M. & Mishra An algorithm is proposed using The algorithm in this Not much improvement been
A. 2014, [12] two conventional task research ensures that the made from FCFS. Some
scheduling algorithm. This jobs are executed with improvement is still needed to
algorithm is based on min-min smaller delays to balance the load between the
and max-min conventional minimize the overall VMs and to minimize the
scheduling techniques. makespan execution time

The aim of the proposed


technique is to minimize the
overall makespan of the jobs
Agarwal A. & Jain S. The proposed algorithm i.e. The proposed Algorithm This Algorithm, prioritized
2014 [13] Generalized Priority Algorithm was found to be better in the job scheduling based on
rank (prioritized) the virtual performance than FCFS MIPS of VM and cloudlets
machines according to their and Round Robin size without considering the
Million Instructions Per Second completion time and load
(MIPS) and the cloudlets balancing.
according to their sizes and it.

The aim of the research was to


proposed algorithm for
scheduling that would help to
schedule task efficiently in
Cloud Computing Environment.

177
The 4th International Conference on Computer Science and Computational Mathematics (ICCSCM 2015)

3. Proposed Multi-Criteria Strategy


Multi-Criteria algorithm (RMK) is a proposed algorithm in
this research where it focuses on job scheduling and resource
load balancing. The algorithm consist three main stages for
job execution. In the first stage, a stack table is formed
containing all the cloudlets information and their execution
time in the available VMs as shown in Figure 1. Meanwhile,
in the second stage a minimum makespan is found by using
three scheduling techniques i.e. min-min, max-min and
suffrage as shown in Figure 2a, 2b and 2c respectively.
Furthermore in stage three, the result from stage two will be
used as parameters to be encoded in to genetic algorithm for
resource load balancing as shown in Figure 3.
This algorithm is believed can enhance job scheduling and
resource load balancing in cloud computing environment. As
a result, this algorithm was tested (simulated) using
CloudSim simulator and found to be of greater impact in
term of job scheduling and resource load balancing.

3.1 Flow Chart of the Proposed Algorithm

Figure 2a. Min-Min algorithm

Figure 1: General format of the proposed algorithm (RMK) Figure 2 b. Max-Min algorithm

178
The 4th International Conference on Computer Science and Computational Mathematics (ICCSCM 2015)

4. Experimental Results
CloudSim simulator is used to test the proposed algorithm.
The setup of this simulation has one datacenter, one
datacenter broker and four Virtual machines (VMs) with ten
cloudlets. However the scheduling policy is space shared,
while the utilizations models of CPU, RAM and Bandwidth
are full utilization models. Meanwhile, the four VMs have
MIPS of 200, 300, 400 and 500 respectively. The ten
cloudlets have different size as 5200, 3000, 2000, 7000,
6800, 6600, 6000, 5800, 5400, 5000 and 4800 respectively.
The results of this simulation shows that this algorithm out
performed FCFS significantly as shown in Table 2 and Table
3 below.

Table 2: Simulation Result using FCFS


Cloudlets VM Datacenter FCFS Time FCFS
ID ID ID Start Taken Finish
Time Time
3 3 2 0.1 80 80.1
2 2 2 0.1 100 100.1
1 1 2 0.1 133.33 133.43
7 3 2 80.1 80 160.1
0 0 2 0.1 200 200.1
6 2 2 100.1 100 200.1
5 1 2 133.43 133.33 266.77
Figure 2c. Suffrage algorithm 4 0 2 200.1 200 400.1
9 1 2 266.77 133.44 400.21
8 0 2 400.1 200 600.1

Summary VMs execution time using FCFS


VM 0: 600.1
VM 1: 400.21
VM 2: 200.1
VM 3: 160.1

FCFS
700

T 600
i
m 500
e
400

T 300
a
k 200
e
n 100

0
V1 V2 V3 V4
Vitual Machine
Figure 3. Genetic Algorithm

Figure 4. FCFS time taken result

179
The 4th International Conference on Computer Science and Computational Mathematics (ICCSCM 2015)

Table 3: Simulation Result using RMK 5. Conclusion


Cloudlets VM Datacenter RMK Time RMK
ID ID ID Start Taken Finish
This research is conducted with the aim of finding an
Time Time
algorithm that is capable to enhance jobs scheduling and
3 3 2 0.1 80 80.1 resource load balancing in cloud computing environment.
0 2 2 0.1 100 100.1 Three scheduling techniques are integrated i.e. min-min,
4 1 2 0.1 133.33 133.43 max-min and suffrage with GA. This integration brings
about the proposed algorithm (RMK) which has three stages
5 3 2 80.1 80 160.1 for execution. Stack table containing execution time of jobs
1 0 2 0.1 200 200.1 mapped to the available VMs, finding minimum makespan
2 2 2 100.1 100 200.1 and load balancing using GA are three stages of this
9 3 2 160.1 80 240.1 algorithm.
The proposed algorithm can enhance jobs scheduling and
8 1 2 133.43 133.33 266.77
resource load balancing as shown in Table 2 and Table 3 in
the previous section.
6 2 2 200.1 100 300.1
7 0 2 200.1 200 400.1
References
Summary VMs execution time using proposed algorithm
(RMK) [1] Google Retrieved from www.google.com/trends
VM 0: 400.1 Visited on 30, May, 2014.
VM 1: 266.77 [2] Erkoc, M. F., Kert, S. B. Cloud Computing for
VM 2: 300.1 Distributed University Campus: A Prototype
VM 3: 240.1 Suggestion. International Conference the Future of
Education.
[3] Kalpeyeva, Z. B., Mustafina, A. K. (2013) IT-
RMK Infrastructure of University based on Cloud
Computing. International Journal of Computer Science
450
Issues, Vol. 10, Issue 5, No. 1, 176 – 179, September
400 2013.
[4] Vouk, M. A. Cloud Computing Issues, Research and
T Implementations. Journal of Computing and IT, 4, 235
350
i – 236, 2008.
m 300 [5] Kart, R., Goldstein, P. and Yanosky, R. Cloud
e Computing in Higher Education.
250 [6] Moothoor, J., Bhatt, V. A Cloud Computing Solution for
Universities: Virtual Computing Lab (2009). IBM
t 200 Corporation 2009.
a [7] White Paper, Introduction to Cloud Computing.
k 150 Dialogic Corporation 9800 Cavendish Blvd., 5th floor
e Montreal, Quebec, Canada H4M 2V9, 2010.
100
n [8] Huth, A. and Cebula, J. The Basics of Cloud
Computing. Carnie Mellon University produced for
50
US-CERT, a government organisation, 2011.
0 [9] Bento, A., Bento, R. Cloud Computing: A new Phase in
V1 V2 V3 V4 Information Technology Management. Journal of
Information Technology Management Volume XXII,
Virtual Machine Number 1, PP 39 – 46, 2011.
[10] Goyal, T. & Agrawal, A. Host Scheduling Algorithm
using Genetic Algorithm in Cloud Environment.
Figure 5. RMK time taken result International Journal of Research in Engineering &
Technology Vol 1, Issue 1, Pp 7 – 12, June 2013.
The results show that if FCFS is used for scheduling the jobs
[11] Gu, J., Hu, J., Zhao, T., Sun, G. (2012). A New
in this scenario, Vm 3 and Vm 2 will be under-utilized and
Vm 0 and Vm 1 will be over utilized. However, the Resource Scheduling Strategy Based on Genetic
implementation of the proposed algorithm (i.e. RMK) Algorithm Cloud Computing Environment. Journal of
achieves significant improvement of the overall time taken Computers Volume 7, No. 1, PP 42 – 52, January,
and resource load balancing between the VMs. 2012.
[12] Katyal, M. & Mishra, A. (2014). Application of
Selective Algorithm for Effective Resource
Provisioning in Cloud Computing Environment.

180
The 4th International Conference on Computer Science and Computational Mathematics (ICCSCM 2015)

International Journal on Cloud Computing: Services [14] Sun, H., Chen, S., Jin, C. & Guo, K. Research and
and Architecture (IJCCSA), Vol. 4, No. 1, February, Simulation of Task Scheduling Algorithm in Cloud
2014. Computing. TELKOMINA, Vol. 11, No. 11, Pp 6664 –
[13] Agarwal, A. & Jain, S. Efficient Optimal Algorithm of 6672, November, 2013.
Task Scheduling in Cloud Computing Environment.
International Journal of Computer Trends and
Technology (IJCTT), Vol. 9, No. 7, Pp 344 – 349,
March, 2014.

181

View publication stats

You might also like