A Novel Resource Selection Method for Cost Optimized Workflow Scheduling with Deadline Constraint using Particle Swarm Optimization for IaaS Cloud
Publication Date : 04/01/2018
Allocation and Scheduling of Cloud resources is an important area in Cloud Computing with minimum execution cost and time. As the cloud is a collection of software and hardware resources and the resources in the cloud are allocated by the cloud service providers, to several users simultaneously based on the users requests, scheduling and optimization play an important role in better resource utilization, faster execution and minimizing the cost incurred in executing these applications. The algorithm used in this paper is based on the meta-heuristic optimization technique, Particle Swarm Optimization (PSO), which is a latest optimization technique. In this paper we present our work which aims to minimize the cost of execution of the workflow while maximizing the makespan to reach the deadline time with a deadline factor and less than or equal to one and hence the title fractional deadline time. The choice of cloud resources is an important factor which controls the cost and makespan of the workflow. The resources used are ordered based on cost and a varied sub-set of resources are considered which resulted in better cost minimization and the features of scalability and elasticity is taken advantage of, for minimizing the cost while meeting the deadlines. This work also compares two billing schemes, namely hourly billing and minute billing. A Java Application with Netbeans IDE was developed using CloudSim framework for implementation of the algorithm. Popular scientific workflows were considered as data for evaluation. The results obtained indicate improved cost minimization with deadline factors than the methods used in the current state-of-the-art methods.
No. of Downloads :