论文部分内容阅读
为了提高计算网格的服务质量(QoS),研究了包含多层调度器的分级分布式网格体系结构中的QoS-aware作业调度问题,提出了一种将计算网格本地调度器作业选择中所使用的作业优先级与全局调度器的作业分配策略相结合的QoS-aware作业调度综合控制策略.建立了一个具有2层调度器的计算网格的随机高级Petri网(SHLPN)模型,并且进行模型精化设计以降低模型求解的复杂性.使用基于SHLPN的性能分析技术进行系统性能评价.数值结果显示QoS-aware作业调度策略能够为高优先级的作业提供较QoS-unaware作业调度策略更好的QoS保证.
In order to improve the quality of service (QoS) of the computational grid, the QoS-aware job scheduling problem in a hierarchical distributed grid architecture with multi-level scheduler is studied. QoS-aware job scheduling integrated control strategy that combines job priority and job dispatch strategy of global scheduler.A stochastic high-level Petri net (SHLPN) model of computing grid with two-tier scheduler is established and implemented Model refinement design to reduce the complexity of the model solution.Using SHLPN-based performance analysis techniques for system performance evaluation.Numerical results show that QoS-aware job scheduling strategy can provide higher-priority jobs than QoS-unaware job scheduling strategy better QoS guarantee.