Performability Evaluation of Resource Scheduling Algorithms for Computational Grids

Yang Qul,Chuang Lin,Yajuan Li,Zhiguang Shan

DOI: https://doi.org/10.1109/gcc.2006.70

2006-01-01

Abstract:Resource scheduling is a fundamental issue in achieving high performance on computational grids. In this paper, we concentrate on parallel resource scheduling problems for hierarchical resource scheduling architecture of computational Grids, and propose a Stochastic Reward Nets (SRN) based performability evaluation framework for evaluating the performability of resource scheduling algorithms for computational Grid with resource unavailability. Within the proposed framework, performability model of hierarchical resource scheduling is constructed, 36 combinations of 6 global job scheduling algorithms and 6 local processor allocation algorithms are specified in the uniform performability model, and performability metrics are derived.