支持资源协同分配的网格任务调度研究

提出通用的网格和任务执行模型,并以此为基础,给出一种支持资源协同分配的任务调度算法。算法通过定义临界资源的概念,改进了传统的列表调度算法。模拟实验结果表明该调度策略更符合网格计算的复杂环境,能得到较短的任务执行时间,并更好的支持不同类型资源的协同分配。

Research on Grid Task Scheduling for Resource Co-allocation

This paper presents models for computational Grids and parallel tasks running in Grid computing environments.Based on these models,a parallel task-scheduling algorithm for resource co-allocation in Grid computing environments is proposed.The algorithm is effective in identifying the critical resource of parallel tasks and scheduling a task as early as possible when all the resources required by the task are available.Simulation results show that shorter execution time and better performance can be gained by utilizing this algorithm,especially in Grid computing Environments.