网格调度的算法研究及改进

优化网格资源调度算法可以提高网格系统执行效率,给任务安排合理的执行顺序和合适的处理器是优化网格资源调度算法需突破的关键技术.文中研究并实现了（Heterogeneous-Earliest-Finish） HEFT[1]算法和新的（Hierarchical Reliability-Driven Scheduling）HRDS算法.采用DAG[2]任务图生成函数,通过对已有HEFT算法进行研究,采用SimGrid为在分布计算环境下进行分布并行应用调度研究提供一个仿真环境,对HRDS算法进行了改进和验证.验证过程中在HRDS算法中加入了可靠性开销作为调度依据,并把算法分为两层调度,本地可靠性驱动调度和全局可靠性驱动调度.两算法的调度结果在SimGrid网格模拟器中仿真调度,仿真成功并且调度结果在可靠性和性能方面HRDS都比HEFT算法要好.

Research and improvement of grid resource scheduling algorithm

The efficiency of grid system can be improved by optimizing the grid resource scheduling algorithm. For optimizing the grid resource scheduling algorithm, the key technology that needs to be broken through is reasonable execution sequence and suitable processor arranged to the task. HEFT algorithm and HRDS algorithm were presented and implemented in this paper. The DAG task graph was adopted to generate function. Based the analysis of existing HEW algorithm, SimGrid was used to offer a simulation environment for distributed parallel application scheduling research under distributed computing environment and to improve and verify the HRDS algorithm. In the verification process, expense reliability is added to the HRDS algorithm as a basis and the algorithm is divided into two levels of scheduling local reliability driven scheduling and global reliability driven scheduling. Scheduling is simulated in the SimGrid grid simulator applying the two algorithms scheduling models. It is shown that the simulation turns out to be successful and both the reliability and the performance of the scheduling based on HRDS algorithm are better than that of HEFT algorithm.