Comparison Of Centralized And Decentralized Scheduling Algorithms Using Gssim Simulation Environment

Various models and architectures for scheduling in grids may be found both in the literature and in practical applications. They differ in the number of scheduling components, their autonomy, general strategies, and the level of decentralization. The major aim of our research is to study impact of these differences on the overall performance of a Grid. To this end, in the paper we compare performance of two specific Grid models: one centralized and one distributed. We use GSSIM simulator to perform accurate empirical tests of algorithms. This paper is a starting point of an experimental study of centralized and decentralized approaches to Grid scheduling within the scope of the CoreGrid Resource Management and Scheduling Institute.