Task scheduling through limited duplication with processor utilization in grid computing system

摘要

In dependent task scheduling algorithms, task duplication is the finest scheduling technique for minimizing the response time of workflow application in grid computing system. When we apply task duplication scheduling algorithm on workflow application, we get shorter schedules (makespan) but it has one limitation that grid node can be overloaded due to duplications of tasks. In this paper we are proposing an algorithm in which we are focusing on three points (1) reducing makespan (2) reducing tasks duplication, and (3) achieving better processor utilization. For this we have suggested an algorithm (TLD-P) which is achieving good results for considered parameters as compared to the existing HLD and EDS-G algorithm.