A Novel Dynamic Priority Based Job Scheduling Approach for Cloud Environment

International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395 -0056
Volume: 04 Issue: 06 | June -2017 www.irjet.net p-ISSN: 2395-0072
© 2017, IRJET | Impact Factor value: 5.181 | ISO 9001:2008 Certified Journal | Page 518
A Novel Dynamic Priority Based Job Scheduling Approach for Cloud
Environment
Er. Shakeel Ahmad1, Er. Imtiyaj Ahmad2, Er. Sourav Mirdha3
1,2M.Tech. Student, Computer Science & Engineering, International Institute of Engineering & Technology,
Samani, Kurukshetra, Haryana, India
3Assistant Professor, Computer Science & Engineering, International Institute of Engineering & Technology,
Samani, Kurukshetra, Haryana, India
---------------------------------------------------------------------***---------------------------------------------------------------------
Abstract - Scheduling of jobs is a foremost and difficult
issue in Cloud Computing. Utilizingcloudcomputingresources
efficiently is one of the Cloud computing service provider’s
ultimate goals. Today Cloud computing is on demand as it
offers dynamic flexible resourceallocationfortrustworthyand
definite services in pay-as-you-use manner, to Cloud service
users. So there must be a provision that all resourcesshould be
made available to demanding users in proficient manner to
satisfy their needs. In this dissertation author has proposed a
new dynamic priority based job scheduling algorithm incloud
computing to optimize the problem of starvation. The priority
in proposed algorithm is based on multiple criteria such as
CPU Resource Requirement, IO ResourceRequirementand JOB
criticality. The proposed model aims to reduce the waiting
time, turnaround time of jobs and to increase the throughput
and CPU utilization of complete system. Acomparisonwith SJF
algorithm in terms of waiting time, turnaround timeandtotal
finish time is performed. Simulation of work has been done on
CLOUDSIM.
Key Words: Cloud Computing, Task Scheduling,
Cloudsim, Shortest Job First
1. INTRODUCTION
Cloud Computing is a term used to illustrate both a platform
and type of application. As a platform it supplies, configures
and reconfigures servers, while the servers can be physical
machines or virtual machines. On the other hand, Cloud
Computing describes applications that are extended to be
accessible through the internet and for this purpose large
data centers and powerful servers are used to host the web
applications and web services [1].
NIST is a well accepted institution all over the world fortheir
work in the field of Information Technology.NISTdefinesthe
Cloud Computing architecture by describing five essential
characteristics, three cloud services models and four cloud
deployment models is shown in figure 1 where layered
architecture is shown [2]
On demand self service, broad network access, resource
pooling,rapidelasticityandmeasuredservicesare5essential
characteristics of Cloud computing which explains there
relation and difference from the traditional computing
system.
.
Fig-1: Cloud computing model given by NIST [2]
2. JOB SCHEDULING
Scheduling is a process of finding the capable resources that
can execute the cloud requests (tasks) at specific times that
satisfy specific performance quality measure such as
execution time minimization,asspecifiedby cloudusers.The
main goal of job scheduling is to achieve a high performance
computing and the best system throughput [3].
Schedulers employ a function that takes into account the
essential objectives to optimize a specific outcome. The
commonly used scheduling reason in a cloud computing
environment is related to the tasks completion time and
resource utilization. The scheduler uses a particular policy
for mapping the tasks to suitable Grid/Cloud resources in
order to satisfy user requirements. However, the bulk of
these scheduling strategies are static in nature. They
produce a good plan given the current state of Cloud
resources and do not take into account changes in resource
accessibility. On the other hand, dynamic scheduling
considers the current state of the system. It is adaptive in
nature and able to fabricate efficient schedules, which
ultimately reduces the completion time of tasks as well as
improves the overall performance of the system [4].
2.1 Starvation
Starvation is a resource management problem where a
process does not get the resources it needs for a long time
because the resources arebeingallocatedtootherprocesses.

Starvation generally occurs in a Priority based scheduling
System where high priority requests get processed first.
Thus a request with least priority may never be processed.
Aging is a technique to reduce starvation in a scheduling
system. It works by adding an aging factor to the priority of
each request. The aging factor must increase the requests
priority as time passes and must ensure that a request will
eventually be the highest priority request [5].
3. CLOUDSIM SIMULATOR
Cloudsim is a new generalized and extensible simulation
framework that enables flawless modeling, simulation, and
experimentation of emerging Cloud computing
infrastructures and management services. Cloudsim has the
following novel features:
1. Support for modeling and instantiation of large
scale Cloud computing infrastructure, including data
centers on a single physical computing node and java
virtual machine
2. Independent platform for modeling data centers,
service brokers, scheduling, and allocations policies
3. Accessibility of virtualizationengine,whichassistin
creation and managementofmultiple,independent,and
co-hosted virtualized services on a data center node
4. Flexibility to switch between space-shared and
time-shared allocationofprocessingcorestovirtualized
services.
Cloudsim is implemented in JAVA language which is based
on the object oriented programming concepts. Class defines
as abstract unit in OOP concepts [6].
4. RELATED WORK
D. Dutta et al. in [7] suggested a genetic algorithm approach
to cost based multi QoS job scheduling. A model for cloud
computing environment has been also proposed and some
popular genetic cross over operators, like PMX, OX, CX and
mutation operators, swapandinsertionmutationareusedto
produce a better schedule. The algorithm assures the best
solution in finite time.
P. Kumar et al. in [7] have discussed various forms of
mapping cluster topology requirements into Cloud
environments to achieve higher reliability and scalability of
application carry out within Cloud resources and enabling
the scheduler to make best use of CPU utilization while
remaining within the constraints imposed by the need to
optimize user Quality of Service (QOS). The focus of the
paper is to provide a dynamic scheduler that aims to
maximize user satisfaction.Thusthejobdetailssubmitted by
the user will include job prioritization criteria i.e. the
allocated budget and the deadline required by the user,
enabling the scheduler to maximize CPU utilization while
remaining within the constraints imposed by the need to
optimize user Quality of Service (QOS).
M. Paul et al. in [9] have proposed scheduling mechanism
which follows the Lexi – search approach to assign the tasks
to the available resources. The scheduled task will be
preserved by a load balancing algorithm that allocate the
pool of task into small partition andthendistributeintolocal
middleware. Cost matrix was generated from a probabilistic
factor based on some most vital condition of efficient task
scheduling such as task arrival, task waiting time and the
most important task processing time in a resource. The
recommended method considered the scheduling problem
as the assignment problem in mathematics here the cost
matrix gives the cost of a task to be assigned into a resource.
Cost had been considered as credit or the probabilistic
measurement thus only the processing time of a job is not
been given importance but the other issues are considered
such as the probability of a resource to be free soon after
executing a task so that it will be available for other waiting
job. Job which has the highest probabilitytogeta resourceas
well as the resource which fits better for a job is assigned in
a manner that one resource get one job at a time. The load
balancing mechanism in the central middleware decreases
the overhead of scheduling on a single middleware by
partitioning the job queue thus scalability issues is well
maintained and making theduplicationofthepartitionedjob
queue ensures the fault tolerant in the cloud since if any of
the client fail then that job could be reassigned into another
client by another local middleware as the local middleware
interact each other for every job updates. The proposed
methodology does not need any complex network
architecture than other job scheduling network architecture
in the cloud.
C.S. Pawar et al. in [10] had put forwarded an algorithm
which considered preemptive task execution and multiple
SLA parameters such as memory, network bandwidth, and
required CPU time. Proposed algorithm dynamically reacts
to fluctuating work load by preempting the current
executing task having low priority with high priority task
and if preemption is not possible due same priority then by
creating the new VM form globally accessible resources. An
achieved experimental resultsshowthatina situationwhere
resource contention is severe proposed algorithm (PBSA )
perform better than CMMS in resource contention situation
and affords better utilization of resources.
A. Tumanov et al. in [11] discussed the need for and an
approach for accommodatingdiversetenant needs, basedon
having resource requests indicate any soft(i.e., whencertain
resource types would be better, but are not mandatory) and
hard constraints in the form of composable utility functions.
They proposed scheduler that acknowledges such requests
that can then maximize overall utility, perhaps weighted by
priorities, taking into account application specifics. Done
Experiments with a prototype scheduler, called alsched,
reveal that support for soft constraints is important for
efficiency in multi-purpose clouds and that composable
utility functions can provide it.
A. Jain et.al. in [12] has criticallyevaluatedtheperformances
of different scheduling algorithms found in literature. The

request time for the three policies applied (Round Robin,
Equally spread current execution load, Throttled Load
balancing) are same which means there is no effect on data
centers request time after changing the algorithms. Thecost
analysis illustrated for each algorithm is calculated in the
experimental work. The cost calculated for virtual machine
usage per hour is same for two algorithms Round Robin,
Equally spread current execution load but Throttled Load
balancing algorithm lessen the cost of usage, so Throttled
Load balancing algorithm works more efficiently in terms of
cost for load balancing on cloud data centers.
5. PROPOSED WORK
Modular representation of proposed approach is shown
below in figure 2.
Fig -2: Modular representation of proposed scheduling
approach
Initialize and scheduler are the major modules. Their
functionality is as follows:
 Functionality of Initialize & Classifier
It is the module that has been generated in Cloudsim which
allocates the five chosen characteristicsvaluetoeachjoband
assigns the initial priority. Initial job pool is created in this
module. The functionality of this module is stated below:
1. It creates the JOBS randomly through CLOUD SIM
2. With each incoming job, some parameters as
associated to all the incoming JOBS
3. Each 1-D array represents the attribute associated
with Job or Process to be executed.
4. IO resource requirement, CPU requirement, Arrival
time, job execution time and job criticality arestatic
parameters.
5. Priority, Wait time, turnaround time andfinishtime
are calculated dynamically.
 Functionality of Scheduler
Scheduler is a module which is responsible for allocation
of jobs to virtual machines on the basis of some priority
value.
1. JOBS in job pool will be initially arranged in
ascending order of their arrival time.
2. Priority for each job is calculated based upon the
values of CPU Requirement, Resource Requirement
and Job Criticality.
3. Jobs are arranged in descending order of priority.
4. Allocations of JOBs to VM are at runtime depending
upon the availability of VM.
5. Searching of VM based upon least execution time is
also done to allocate the unassigned jobs.
6. Priority of unassigned jobs is again calculated and
incremented by 1 if wait time of job exceeds wait
threshold value.
7. Non-Preemptive dynamic Scheduling is performed
“Starvation Optimizer Scheduler” is dynamic algorithm
based upon the priority assigned to each task.Thealgorithm
starts its operation by first creating the job pool where in
jobs are created and five characteristics (Arrival time, CPU
execution time, CPU requirement, IO resource requirement
and job criticality) are associated with each job. These
characteristics form the basistocalculateorassigntheinitial
priority for each job. Once the priority of jobs is calculated,
jobs are sorted in descending order of priority. Higher
priority jobs are assigned to virtual machines. For the
remaining unassigned jobs search for VM having least
execution time is done before allocation of jobs. The
assignment of job to VM depends upon their priority. If the
waiting time of job exceeds the wait thresholdvalue,priority
of job is incremented by one. Finally wait time, turnaround
time and finish time for each job is calculated. Following
formulas are used for calculation:
Wait Time: = Start Time – Arrival Time
Turn Around Time: Finish Time –Start Time
Throughput: (CPU Clocks used in process
execution)/ (Total Clocks)*100
Algorithmic form of starvation optimizer scheduler is as
follows:
1 Enter the number of jobs to be executed.
2 While (J!=NULL) // J is Job Pool
3 For each Job (ji) ϵ J
Initialize arrival_time, execution_time,
cpu_requirement, IO_requirement & Job_criticality.
End For
4 Arrange all Jobs ji in the ascending order of arrival
time.
5 For each Job (ji)ϵ J
Calculate the Job_Priority (Pi) based upon
cpu_requirement,IO_requirement & Job_criticality.
End For
6 Arrange jobs in descending order of priority.
7 Allocate high priority jobs to VM for execution.
a) Compare job_wait_time (WTi) with
wait_threshold

b) If job_wait_time (WTi) > wait_threshold THEN
Increment the priority for job (ji) by 1
End for
9 Search VM ( Vk) having least execution time.
Allocate high priority jobs to VM( Vk) and calculate
wait time and turnaround time by using following
formulae:
a) Wait Time = Start Time (ji)– Arrival Time( ji)
b) Turn Around Time= Finish Time(ji) –Start
Time(ji)
c) Update the status of Job to complete.
END While
6. RESULTS & ANALYSIS
6.1 Simulation configuration
A simulation program is implementedinJAVA language with
the help of Cloudsim tool kit to optimize the starvation
problem in cloud environment. Simulation is implemented
under following set of assumptions:
 Type of Scheduling: Non pre-emptive and Dynamic.
 For same priority jobs FCFS scheduling policy will
be used.
 Highest priority value is 3 and lowest is 1.
 Waiting threshold value is 40 second.
Random job pool of six jobs is created. For each job - Arrival
Time, CPU Clock, CPU Requirement, Resource Requirement
and Job Criticality is provided as input characteristic.
Table 1 shows valuesforthe abovecharacteristicsassociated
with each job
.
Table -1: Input Data Set for SOS Algorithm
Proc
ess
Arriv
al
Time
CPU
Execu
tion
CPU
Require
ment
Resource
Require
ment
JOB
Critica
lity
P0 2 10 3 3 3
P1 2 20 1 1 1
P2 2 30 2 2 2
P3 3 25 1 2 3
P4 3 20 3 2 3
P5 3 10 3 5 2
6.2 Results
Scheduler will assign the jobs to VM and for each job finish
time, wait time and turnaround time is calculated as output.
The obtained output characteristics values are shown in
table 2.
Table -2: Output Data Set for SOS Algorithm
Process Finish
Time
Turnaround
Time
Wait
Time
Priority
P0 12 10 0 2
P1 22 20 0 1
P2 32 30 0 1
P3 49 25 21 3
P4 43 20 20 3
P5 23 10 10 2
Evaluation summary forall thejobscomprisingoftotal finish
time, CPU Utilization, throughput, average turnaround time
and average waiting time is presented in table 3.
Table 3: Evaluation Summary for SOS
Parameters Values
Total Finish Time 49
CPU Utilization 0.87
Throughput 40
Average Turnaround Time 19.16
Average Waiting Time 8.5
Quantitative analysis of SOS algorithm is presented in figure
3. For each job- finish time, turnaround time and waittimeis
depicted with different colors. Finish time is shown in blue
color, turnaround time in red color and wait time in green
color. As for the first three jobs, the wait time is zero, so the
weight time bar is not figured.
Fig -3: Quantitative Analysis of SOS Algorithm
Quantifying results in figure 4 shows that by dynamically
increasing the priority of jobs,averagewaitingtimeandtotal
finish time of the complete system is reducedapproximately
to 23% and 4% respectively. Hence problem of starvation is
optimized in “Starvation Optimizing Scheduler” algorithm.

Fig -4: Quantitative Analysis of SOS Algorithm
7. CONCLUSIONS
Job scheduling problem is important and challenging issue
in Cloud Computing. Utilizing cloud computing resources
proficiently and gaining the highest profits with job
scheduling system is one of the Cloud computing service
providers’ ultimate goals.
Research done earlier in this area was focused on mapping
of tasks to machines efficiently but still problem of
starvation persists. So to resolve this issue major focus of
this paper has been put on optimizing the starvation. An
algorithm “SJF” mainly suffers from this problem. New
algorithm is generated “Starvation OptimizingScheduler”
which aims to reduce the starvation. Following objectives
have been met satisfactorily which are stated below:
 Jobs are allocated to VM’s dynamically at run time.
 The average waiting time, average turnaroundtime
and total finish time of jobs are reduced.
 Starvation problem is optimized.
Also this work can be extended in future in the following
way:
1 In this work, author has input the jobs only once
under different arrival time specification, but no
work is defined for the job input during the job
execution. In future, work can improved by
including the anytime participation of userinterms
of job input.
2 In this work, jobs are defined in non-preemptive
way, but in future thee technique of preemptioncan
be used for allocating resources to jobs.
REFERENCES
[1] A. Jain and R. Kumar, “A TaxonomyofCloudComputing,”
International Journal of Scientific and Research
Publications. vol. 4(7), Jul. 2014, pp. 1-5.
[2] G. Brunette & R. Mogull, “Security guidance for critical
areas of focus in cloud computing v2. 1,” Cloud Security
Alliance, 2009, pp. 1-76.
[3] R. Buyya, C.S. Yeo, S. Venugopal, J. Broberg & I. Brandic,
“Cloud computing and emerging IT platforms: Vision,
hype, and reality for delivering computing as the 5th
utility”. Future Generationcomputersystems, vol.25(6),
2009, pp. 599-616.
[4] H. Topcuoglu, S. Hariri & M. Y. Wu, “Performance-
effective and low-complexity task scheduling for
heterogeneous computing,” Parallel and Distributed
Systems, vol. 13(3), 2002, pp.260-274.
[5] M. Rahman, S. Venugopal & R.Buyya,“Adynamiccritical
path algorithm for scheduling scientific workflow
applications on global grids,” In e-Science and Grid
Computing, IEEEInternational Conference,2009,pp.35-
42.
[6] R. Buyya, R. Ranjan & R.N. Calheiros, “Modeling and
simulation of scalable Cloud computing environments
and the CloudSim toolkit:Challengesandopportunities,”
High Performance Computing & Simulation, HPCS'09.
International Conference, 2009, pp. 1-11.
[7] D. Dutta & R.C. Joshi, “A genetic: algorithm approach to
cost-based multi-QoS job schedulingincloudcomputing
environment”. In Proceedings of the International
Conference & Workshop on Emerging Trends in
Technology, 2011, pp. 422-427.
[8] P. Kumar, N. Nitin, V. Sehgal, D. S. Chauhan & M.
Diwakar, “Clouds: Concept to optimize the Quality of
Service (QOS) for clusters,” In Information and
Communication Technologies (WICT), 2011, pp. 816-
821.
[9] M. Paul, D. Samanta, & G. Sanyal, “Dynamic job
Scheduling in CloudComputingbasedonhorizontal load
balancing,” International Journal of Computer
Technology andApplications (IJCTA), vol.2(5),2011,pp.
1552-1556.
[10] C.S. Pawar & R.B. Wagh, “Priority Based Dynamic
resource allocation in Cloud Computing,” In Cloud and
Services Computing (ISCOS), 2012, pp. 1-6.
[11] A. Tumanov, J. Cipar, G.R. Ganger & M.A.
Kozuch,”Algebraic scheduling of mixed workloads in
heterogeneous clouds,” InProceedings of theThirdACM
Symposium on Cloud Computing, 2012, pp. 25-30
[12] A. Jain and R. Kumar, “A Comparative Analysis of Task
Scheduling Approaches for Cloud Environment,”
International Conference On Computing for Sustainable
Global Development, 2016, pp. 2602-2607.

A Novel Dynamic Priority Based Job Scheduling Approach for Cloud Environment

More Related Content

What's hot (19)

Similar to A Novel Dynamic Priority Based Job Scheduling Approach for Cloud Environment (20)

More from IRJET Journal (20)

Recently uploaded (20)

A Novel Dynamic Priority Based Job Scheduling Approach for Cloud Environment