Comparison of Average Completion Time and Makespan for Various Task Scheduling Techniques

Abstract

Cloud computing is becoming a trending and important technology that everyone is using globally on a large scale. It allows users to get access to resources (Storage and computational) online over the internet. With the increasing number of users, maintaining the Quality of service is challenging. The jobs are becoming more and more intensive and hence designing and implementing the correct task scheduling algorithm is very necessary to fulfill the increasing demand of users. An algorithm is needed to be designed to map the request of users to a particular Virtual Machine to utilize the resources of virtual machines efficiently. In this paper, we have observed from the results that better makespan and optimum or efficient resource usage are required for both customers and cloud service scheduling algorithms, which are considered suitable for large-scale systems. In this paper, the performance of 5 scheduling algorithms - FCFS, SJF scheduling, Round Robin scheduling, general priority algorithm, and Particle Swarm Optimization (PSO) is analyzed and compared. The performance of the various task scheduling algorithms has been carried out on the Cloudsim simulator. It has been observed from results that PSO outperforms the other compared task scheduling techniques and gives 48.3% less makespan as compared to FCFS.