A Novel Decentralized Fuzzy Based Approach for Grid Job

Abstract

In this paper with the aid of fuzzy theory we present a new method for scheduling on Grid system. Grid computing is a technology to meet the growing computational requires. In fact grid computing is one of the most popular types of distributed system. Its aim is to produce an enormous, autonomous and effective virtual machine, and it is produced by collecting different nodes with the aim of sharing their data and computational power. This paper follows the identification of grid scheduling with the help of fuzzy theory and seeking to present a new method for grid scheduling with respect to exiting obstacles. In our method we use the intermediate load of nodes of each clusters, the average of computing power which determines the node premiership and job premiership as the input parameters of fuzzy system, and regarding to the output value of fuzzy system the suitable nodes determines. We evaluate the performance of our method with some grid scheduling methods. The results of the experiments show the efficiency of the proposed method in term of makespan and Standard deviation of the load of clusters