Performance comparison of process allocation schemes depending upon resource availability on grid computing environment

Abstract

Improvements in the performance of end-computers and networks have recently made the construction of a grid system over the Internet feasible. A grid environment comprises many computers, each having a set of components and distinct performance, that are shared among many users and managed in a distributed manner. Thus, it is important to focus on a situation in which the computers are used unevenly due to decentralized management by different process schedulers. In the present study, as a preliminary theoretical investigation of the effect of such features on the performance of process schedulers, the average execution time of a long-lived process in the process allocation scheme employed in the decentralized environment is analytically derived using M/G/1-PS queues. The impact of the distribution of CPU utilization on the performance of these schemes is also investigated assuming CPU utilization for each computers in a grid environment follows a probability distribution function. © Springer-Verlag Berlin Heidelberg 2003.