You are on page 1of 29

JOB SCHEDULING ALGORITHMS IN CLOUD COMPUTING

(WITH SPECIAL IMPORTANCE ON ROUND-ROBIN ALGORITHM)

By:
Sendash Pangambam
Dept. of Computer Science
Banaras Hindu University
Varanasi, India
OVERVIEW:
1. CLOUD COMPUTING
2. CHARACTERISTICS
3. WORKING MODELS
4. SCHEDULING
5. SOME SCHEDULING ALGORITHMS AND COMPARISIONS
6. OBSERVATIONS ON RR ALGORITHM
7. CONCLUSIONS
8. REFERENCES
CLOUD COMPUTING???
Cloud: Something at remote location which can provide services
over network/internet.

Cloud computing:
A type of computing consisting of inter-connected and virtualized
computers that are dynamically provisioned, based on service-
level agreements established between the service provider and
consumers.
It delivers infrastructure, platform, and software that are made
available as subscription-based services.
CHARACTERISTICS:
1. On demand self service: Capable to provision computing
capability as needed automatically.
2. Broad network access: Capability to access by thin or thick
client platforms.
3. Resource pooling: Computing resources are pooled to serve
multiple consumers.
4. Rapid elasticity: Capable to provision resources dynamically.
5. Measured service: Resource usage should be monitored,
controlled; transparency for provider and consumer about
the utilized service.
WORKING MODELS:
 Deployment Models: Type of access
 Public (Service open for public use, less secure)

 Private (Service solely for a single organization)

 Community (Shares infrastructure between organizations


with common concerns, reduce cost)

 Hybrid (Private + Public cloud, critical and non critical


activities)
WORKING MODELS:
 Service Models: Reference model the cloud is based
 IaaS (Provides the necessary hardware and software
upon which a customer can build customized
environment)

 PaaS (Provides online software development


platform, developer can develop their own
applications)

 SaaS (Software applications are provided by cloud


service provider)
Cloud Service Models
Software as a Platform as a Infrastructure as a
Service (SaaS) Service (PaaS) Service (IaaS)

SalesForce CRM

LotusLive

Google
App
Engine

Adopted from: Effectively and Securely Using the Cloud Computing Paradigm by peter Mell, Tim
7
Grance
SCHEDULING:
 A mapping mechanism from users’ tasks to the appropriate
selection of resources and its execution.
 One of the challenging issue.
 To spread the load on processors and maximize their
utilization, minimize the execution time without affecting
cloud service.
 Involves finding out a proper sequence in which jobs can be
executed under transaction logic constraints.
 Aims at high throughput with least execution time.
SCHEDULING STEPS
1. Resource discovering and filtering:
Datacentre Broker discovers the resources present in the
network system and collects status information related
to them.
2. Resource selection:
Target resource is selected based on certain parameters
of task and resource.
3. Task submission :
Task is submitted to resource selected.
EXISTING JOB SCHEDULING ALGORITHMS:
FCFS:

 One of the simplest algorithms.


 Tasks are executed in the order they arrive.
 Tasks are executed in FIFO manner without any
other preference.
 Long waiting time.
EXISTING JOB SCHEDULING ALGORITHMS:
SJF:

 Selects task with least execution time.


 Jobs are queued with the least execution time
placed at first, task with longest execution time at
last.
 Low utilization of available resource, long waiting
time.
EXISTING JOB SCHEDULING ALGORITHMS:
PSA:

 Tasks done based on some priority.


 Highest priority done first lowest priority at last.
 Starvation of process.
EXISTING JOB SCHEDULING ALGORITHMS:
GA:

 Greedy approach.
 Genetics as problem solving model.
 Random searching method using some parameters.
EXISTING JOB SCHEDULING ALGORITHMS:
RR:

 Scheduler maintains a circular queue of the tasks.


 Each task is executed for a slice of time.
 If executed within the slice of time, then removed.
 If not, added to the tail of the queue and wait for next
round.
 Criteria: Context switch, throughput, TA time, WT, RT.
 Utilize resource in balance order, VM fairness.
COMPARISION:
Algorithm Complexity Allocation Waiting time Type of the system

FCFS It is the simplest scheduling Allocated in the order in which More Suitable for Batch system.
algorithm. processes are arrived.

SJF It is difficult to understand code. Allocated to the process with least Less than FCFS Suitable for Batch System.
burst time.

PSA It is difficult to understand. Higher priority processes can run Lesser Suitable for both Batch and time
first. sharing systems.

GA Complexity of GA depends on the Greedy algorithm which picks the Waiting time is less. Deals with problems where search
task to be scheduled. best job. space is large.

RR Performance depends on the time Preemption takes place after a More than all. Suitable for time sharing systems.
slice (quantum). fixed interval of time.
OBSERVATIONS ON RR ALGORITHM:

 Round robin scheduling algorithm is implemented with the


help of CloudReports.
 CloudReports is a Cloudsim based GUI tool.
 One data centre with five virtual machines are used for our
experiment.
 The parameter values used in the simulation are given in the
next slide.
PARAMETER VALUES USED:
Parameter Value

No. of data centre 1

No. of host 1

No. of processing units 4

Processing capacity(MIPS) 9600

Storage capacity 11 TB

Total amount of RAM 40 GB

VM Image size 1000

VM Memory size 512

VM Bandwidth 100000

Data centre-Architecture x86

Data centre-OS Linux

Data centre-VMM Xen

Data centre-No. of machines 5

Data centre-Allocation policy Single threshold

Data centre-No. of hosts 1

Data centre-VM policy Time shared

Processing cost per second 0.1

Memory cost per MB 0.05

Storage cost 0.001

Bandwidth cost per MB 0.1

No. of customers 1

Cloudlets sent per minute 50

Average length of the cloudlet 40000

Average cloudlet’s file size 300

Average cloudlet’s output size 300

Customer-Broker policy Round robin

Customer- No. of VM 5
OVERVIEW OF THE SIMULATION
OVERALL RESOURCE UTILIZATION ON THE DATACENTRE:
OVERALL POWER CONSUMPTION ON THE DATACENTRE:
NO. OF VMS ALLOCATED SUCCESSFULLY ON THE DATA CENTRE BY EACH CUSTOMER
NO. OF SUCCESSFULLY EXECUTED CLOUDLETS ON THE DATACENTRE BY EACH CUSTOMER
COST GENERATED BY EACH CUSTOMER BASED ON THEIR RESOURCE UTILIZATION
OVERALL RESOURCE UTILIZATION OF THE CUSTOMER
NO. OF CLOUDLETS SUCCESSFULLY EXECUTED BY THE CUSTOMER ON EACH CUSTOMER
AMOUNT OF TIME EACH CUSTOMER'S CLOUDLET TOOK TO BE EXECUTED
CONCLUSION:

 Not a single scheduling algorithm provides superior


performance.

 Depends on tasks to be executed based on its ability to ensure


good QoS maintain fairness by fairly distribute the available
resources among all the jobs and respond to the constraints of
the users.
REFERENCES:
 Cloud Computing Made Easy by Cary Landis and Dan Blacharski

 Comparative Study Of Scheduling Algorithms In Cloud Computing Environment by Isam Azawi Mohialdeen College
of Information Technology, University Tenaga Nasional

 Round Robin Approach for VM Load Balancing Algorithm in Cloud Computing Environment by Nusrat Pasha, Dr.
Amit Agarwal, Dr. Ravi Rastogi Sharda University, Noida.

 Efficient and Enhanced Algorithm in Cloud Computing by Tejinder Sharma, Vijay Kumar Banga, International Journal
of Soft Computing and Engineering (IJSCE) ISSN: 2231-2307, Volume-3, Issue-1, March 2013.

 International Journal of Science and Research (IJSR), India Online ISSN: 2319-7064 Volume 2 Issue 6, June 2013

 Scheduling Virtual Machines for Load balancing in Cloud Computing Platform by Supreeth S, Shobha Biradar
Department of Computer Science and Engineering, Reva Institute of Technology and Management Yelahanka,
Bangalore, Karnataka, India

 Analysis of variants in Round Robin Algorithms for load balancing in Cloud Computing by Pooja Samal, Pranati
Mishra International Journal of Computer Science and Information Technologies, Vol. 4 (3) , 2013.

 Analysis of Job Scheduling Algorithms in Cloud Computing by Rajveer Kaur, Supriya Kinger International Journal of
Computer Trends and Technology (IJCTT) – volume 9 number 7 – Mar 2014

 http://cloud.cio.gov/topics/cloudcomputingservicemodels

 en.wikipedia.org/wiki/Cloud_computing
THANKS

You might also like