Optimization of FCFS based Resource provisioning

algorithms for Cloud Computing
Guided by Prof. Lakshmi Kurup
Aditya Marphatia

Aditi Muhnot

Dept. of Computer Engineering
DJSCOE
Mumbai, India
adityamarphatia@gmail.com

Dept. of Computer Engineering
DJSCOE
Mumbai, India
aditimuhnot@gmail.com

Tanveer Sachdeva

Esha Shukla

Dept. of Computer Engineering
DJSCOE
Mumbai, India
tanveersachdeva@gmail.com

Dept. of Computer Engineering
DJSCOE
Mumbai, India
eshashukla91@gmail.com

Abstract— In our project, we propose an optimized version of
the FCFS scheduling algorithm addresses these major challenges
of task scheduling in cloud. The incoming tasks are grouped on
the basis of task requirement like minimum execution time or
minimum cost and prioritized (FCFS manner). Resource
selection is done on the basis of task constraints using a greedy
approach. The proposed model will be implemented and tested
on simulation toolkit. We intend to create a module depicting the
normal FCFS algorithm in comparison to our optimized version
algorithm for resource provisioning in the cloud.
Index Terms—Cloud computing, FCFS, module, resource,
scheduling

efficiently and to their best capacity so that resource
potential is not left unused.
Much of the research so far has been focused on Cloud
security and data protection aspects, therefore the efficient
provisioning of resources and necessity of cloud services to
be cost effective is often neglected. In our project, we
propose an optimized version of the FCFS scheduling
algorithm which addresses these major challenges of task

I. INTRODUCTION

scheduling in cloud. The incoming tasks are grouped on the

Cloud computing has emerged as a popular computing

basis of task requirement like minimum execution time or

model to support on demand services. It is a style of

minimum cost and prioritized (FCFS manner). Resource

computing

are

selection is done on the basis of task constraints using a

delivered as a service to external customers using Internet

greedy approach. The proposed model will be implemented

technologies. Scheduling in cloud is responsible for

and tested on simulation toolkit. We intend to create a

selection of best suitable resources for task execution, by

module

taking some static and dynamic parameters and restrictions

comparison to our optimized version algorithm for resource

of tasks’ into consideration. The users’ perspective of

provisioning in the cloud.

where

massively

scalable

resources

efficient scheduling may be based on parameters like task
completion time or task execution cost etc. Service
providers like to ensure that resources are utilized

depicting

the

normal

FCFS

algorithm

in

The cost based tasks are prioritized on the basis of III. will be using various While current cloud systems are beginning to offer current resource provisioning strategies. This is mentioned have certain flaws on which work could appreciable as tasks with higher profit can be be done in this direction to extract the advantageous executed on minimum cost based machine to points of these algorithms and come up with a better give maximum profit. the task list is rearranged the preference of the FCFS scheduling. components on the basis of certain behaviour or Resource provisioning algorithms have certain attribute. 3. solution that tries to minimize the drawbacks of resultant algorithm. minimum time constraint first. is a new and 1. PROBLEM DEFINITION WITH SCOPE OF PROJECT • The algorithm devised by us. efficiency of traditional protection mechanisms are being reconsidered as the characteristics of this innovative deployment model can differ widely 2. Deadline Constrained unfamiliar paradigm for computer application 2. if employed to combine several the cloud being under threat. who are accustomed to working with a fixed set of resources they own. Grouping. There are the two ways of choosing the deciding parameters for the utility-like provisioning of services. it There can be better more-costly implementable determines the order of task scheduling based on resource provisioning algorithms such as the the parameters undertaken for its computation. However it with tasks arranged in ascending order of requires many more factors to be considered and deadline in order to execute the task with the cost rises up. In terms of task scheduling. Greedy Allocation: Greedy algorithm is suitable for dynamic heterogeneous resource environment connected to the scheduler through . By task grouping in cloud it is meant budget constraints and the implementation of the that tasks of similar type can be grouped together cloud on a large-scale is a very costly affair.II. In the proposed framework tasks are grouped on the basis of Cloud security has been a major issue uptil now. In probability dependent priority used by Ye Hu over the present framework. • Task Grouping: Grouping means collection of associated. and then scheduled collectively. Hence. Prioritization: Priority determines the importance of the element with which it is from those of traditional architectures. Cost Constrained developers and users. Outline of the Proposed Algorithm: Resource cost is an important consideration while using cloud resources for scientific computing as • • 1. well. however. reduces the cost-communication ratio. they are given more priority in scheduling sequence. constraint which can be deadline or minimum with the privacy of the databases being stored on cost. This. PROPOSED SYSTEM The different resource provisioning algorithms task profit in descending order. The effectiveness and tasks. provisioning of resources has to be controlled by implementation of the proposed algorithm: the end users.

Minimum Cost Based: The resource with minimum cost is selected and tasks are scheduled on it until its capacity is supported. Such algorithms are called greedy because while the The various factors to be kept in mind when studying optimal solution to each smaller instance will the feasibility of the cloud are: provide an immediate output. We will to solve the job scheduling problem. FEASABILITY OF PROPOSED SYSTEM issues minimise and when it is needed. 3. Technological and System feasibility: An error-free simulation analysis is very important point that needs to be considered IV.homogeneous communication environment. risks and gaps. provide the most obvious benefit. Cloud Feasibility: An understanding of where you Greedy approach is one of the approaches used are and what your objectives mean. easy-to- cloud component.Customers need only to buy what they require today rather than buying an IT solution that they will have to grow into overtime and then provision extra resource as This being a cloud computing project. But still this should not be a major concern. resiliency delivered to the customer. the algorithm 1. Comprehensive analysis not only implement solutions to complex. the client and the cloud service provider. feasibility to on maximising the utilisation of computing greedy algorithm is used with aim of minimizing ii. completion time of tasks the turnaround task of individual tasks. The ACID properties of the cloud database need to be satisfied. Deadline Constrained Based: To improve the for downtime enabling SLA based services to be Realization i. . specifications Different Parameters for Greedy Approach 2. this project can face obstructions as it deals with interfacing with heterogeneous environments. The Algorithm can be implemented using Java. the major will be in the efficient implementation of the cloud environment and in the simulation of the proposed algorithm to have better utilization of available resources. Flexibility . multi-step identifies benefits. but also determines problems by deciding which next step will the financial impact on the project. doesn’t consider the larger problem as a whole. determine what hardware and applications can be optimized through virtualization. Efficiency – The underlying technology focuses power and disk storage resulting in less wasted resource. High availability – designed to the highest On cloud side. Operational Feasibility: The algorithm that is proposed should have a considerable positive impact on the resource provisioning to both parties i.e. a fundamental Greedy algorithms look for simple. Economic Feasibility: Economics of this project need to be understood from the development of the cloud environment perspective.

of Computer Engineering for providing us with the resources and infrastructure to undertake this project. B. Additionally. AN The CloudSim toolkit supports both system and behavior modeling of Cloud system components such as data centers.. M. The different resource provisioning algorithms are compared using this simulator software and it is proved that the proposed algorithm is better in resource utilization than other algorithms.Currently there is an existing toolkit that depicts the different resource provisioning algorithms called as CloudSim: an extensible simulation toolkit that enables modeling and simulation of Cloud computing systems and application provisioning environments. A. M. ACKNOWLEDGMENT We wish to thank Prof. pp 7-12. • [16] ADABI. . 2011. V. “A CONTINUES DOUBLEAUCTION METHOD FOR RESOURCE ALLOCATION IN ECONOMIC • Grids”.. ASSUMPTION Our work is based on the assumption that finegrained allocation and pricing of resources is possible for each virtual environment. A. ON GRID COMPUTING (GRID 2008) • 85-94. while evaluating the efficiency of the proposed algorithm in comparison with other algorithms assumes that he jobs arrival is Uniformly Randomly Distributed to get generalized scenario.. CLOUD COMPUTING RESOURCEALLOCATION MECHANISM RESEARCH BASED ON REVERSEAUCTION . 2010 .IT existing cloud environments. H. A. PP. JAVA AGENT DEVELOPMENT FRAMEWORK..[11]GROSU. REFERENCES • IEEE/ACM INT. TORK LADANI . LADANI.SNASEL.. V. H. CONF.FUTURE GNERATION COMPUTER SYSTEMS VOL 26 . TEYMOURI. and recent research points to the progression of clouds in such a direction. B. RAHMANI.HTTP://JADE. International Journal of Applicationvol 43. 2011 . “A CONTINUOUS DOUBLE AUCTION METHOD FORRESOURCE ALLOCATION IN COMPUTATIONAL GRIDS”. D. ZAMANIFAR.. This is not true for the • XIAOHONG • [15] I.MARKET-BASED GRID RESOURCE ALLOCATION USING NEW NEGOTIATION MODEL”. DAS. 2004. 2008.. JOURNAL OF NETWORK ANDCOMPUTER APPLICATIONS. The user of an instance can request for a different CPU cycle percentage or memory allocation at any point during its execution. [17]S. IEEESYMPOSIUM ON COMPUTATIONAL INTELLIGENCE INSCHEDULING. Thus. 16TH IASTED INTERNATIONAL CONFERENCE ON PARALLEL AND DISTRIBUTED COMPUTING AND SYSTEMS. A.. Narendra Shekokar and our guide Prof.ABRAHAM. 2004.. “AUCTIONBASED RESOURCE ALLOCATION PROTOCOLS IN GRIDS”. K.. 2012. • [12] IZAKIAN. ENERGY PROCEDIA VOL 13. RAHMANI. 2012. 2009. MOVAGHAR.. we assume that CPU cycles and memory can be shared in a fine-grained fashion between different instances.. but is consistent with WU AT LA. A. AUCTION FOR RESOURCE ALLOCATION IN COMPUTATIONAL GRIDS . virtual machines (VMs) and resource provisioning policies... BEIGY. • [20] JADE. 29-35. 2009.PP 228–.T. 2012. Computer . Lakshmi Kurup for their continued encouragement and support and the Dept. FUJIWARA . • [13]H.PP 736–741 the utility vision of computing. S.CSELT. A. ABRAHAM. • IEEE 10TH ANNUAL INTERNATIONAL SYMPOSIUM ONAPPLICATIONS AND THE INTERNET. IZAKIAN. APPLYING DOUBLE-SIDED COMBINATIONALAUCTIONS TO RESOURCE ALLOCATION IN CLOUDCOMPUTING .