Two stage strategy of job scheduling in grid environment based on the dynamic programming method

Abstract

The problem of the effective usage of the grid environment for solving different types of computing tasks of large dimension is researched in the paper. We study the problem of optimal task scheduling at an affordable set of resources on the one hand and the equitable distribution of resources between the tasks that come into the input queue of a centralized workflow management system, on the other hand. Two stage strategy of task scheduling in grid environment that takes into account user-defined QoS requirements, structural features and execution dynamicity of the task is presented. The dynamic programming method application to the workflow scheduling problem is proposed in the paper and the effectiveness evaluation experimental results of the proposed decision are given