基于计算期望和信誉度的网格资源调度模型

Grid resource scheduling model based on computing expectation and credit

目的研究网格经济模型下资源调度的问题。方法分析以往的资源调度策略,提出了一种基于计算期望和资源信誉度相结合的网格资源调度模型。结果通过用户设定各因素的重要性,由计算期望分析器将各个因素量化,再由调度器根据量化结果,将任务放到相应的队列中,各队列调用不同的调度算法;同时每次任务完成后,用户的满意度又直接影响资源的信誉度。结论该模型实现了高效灵活的网格资源调度,达到了资源最优分配的目的。

Aim To study grid resource scheduling under the theory of grid economy model. Methods Based on analysis of normal resource scheduling strategy, a computing expectation and credit of the grid resource scheduling model is presented. Results Users can set the importance of each factor, computing expectation analyzer quantifies the various factors, dispatcher puts the task into corresponding tasks queue according to the quantifiable results, then, each queue uses different scheduling algorithm. At the same time, each task is completed, the user＇s satisfaction has impact on the credit of resources directly. Conclusion The results show that the model achieves flexible and efficient grid resource scheduling and optimum resources allocation.