Range-Suffrage Algorithm for Grid Task

Abstract

Scheduling is a fundamental process for grid computing systems. Its goal is to map user tasks to suitable resources for execution. The major part of a grid scheduler is to decide which resource is suitable for each task, depending on a scheduling algorithm. Many scheduling algorithms have been designed for reaching optimality. The Suffrage algorithm has shown a superlative performance over most meta-task scheduling algorithms regarding resources selection. However, providing a full power use of resources is still a challenge. In this paper, a new heuristic algorithm is proposed. It aims to maximizing the resource utilization and minimizing the makespan. Its decision is based on detecting the maximum average value of completion times among certain tasks. These tasks are selected depending on their suffrage values. The task having the maximum average is then assigned to the resource with the minimum completion time. Experimental results show that the proposed algorithm outperforms other algorithms in terms of utilization and makespan.