摘要
针对网格环境中应用程序常为复杂的计算密集型的并行分布式应用程序,提出了一个新的基于复制和插入的启发式任务调度算法(duplication-and-insertion-based scheduling,DIBS),可以同时执行多个应用程序,利用决定路径对任务进行排序,缩短了应用程序总的执行时间,该算法还平衡了处理器间的负载.实验结果表明,该算法更加符合网格的复杂环境,能够更好地满足不同用户的实际需要.
Focusing on the problem of complex computation crowded parallel distributional application procedure in the grid, a new duplication-and-insertion-based heuristic task scheduling algorithm, namely, duplication-and-insertion-based scheduling (DIBS) is proposed, which executes more applications simultaneously, and sorts using the decisive path of the DAG to reduce the makespan of all applications. The proposed algorithm balances the load on each machine. This algorithm is more suitable for the complex environment and assures the QoS of tasks more effectively in the grid.
出处
《计算机研究与发展》
EI
CSCD
北大核心
2008年第z1期21-25,共5页
Journal of Computer Research and Development
基金
教育部科学技术研究重点项目(205014)
河北省教育厅科研计划基金项目(2006143)
