【摘要】 文中以星型网格应用环境中独立任务周期性调度的研究成果为基础,通过分析网格平台分别处于非饱和态、临界态以及冗余态时的性质,提出了拓扑结构优化方法——静态优化与动态优化,给出了一种星型网格计算能力与资源利用率的度量方法;在静态优化方面,解决网格拓扑结构最优化的判断问题,给出组建最优化星型网格的方法;在动态优化方面,研究了网格拓扑结构动态变化时节点的准入策略;还根据网格计算能力与动态负载的匹配情况,提出了网格拓扑结构动态优化算法.最后给出了网格平台的扩展策略.更多还原

【Abstract】 Based on the periodic scheduling research of independent tasks on the star-based grid,a new topology optimization method including static and dynamic optimizations is proposed for the star-based grid by analyzing different features of grid platforms respectively in unsaturated,critical and redundant states,and a method to measure the computing capability and resource utilization of the star-based grid is put forward.Through the static optimization,the optimal grid topology structure is identified,and a method to build the optimal grid platform is presented.Through the dynamic optimization,the admittance strategies of a new node to the grid platform with the dynamic topology structure are discussed.Then,by mapping the computing capability to the dynamic loads,a dynamic optimization algorithm of grid topology structure is raised.Finally,the expansion strategy of grid platform is devised.更多还原