气象环境下的云计算调度算法

Cloud Computing Scheduling Algorithm in Meteorological Environment

针对气象计算的特点,提出气象计算的云模型,在这个模型之上,提出气象云计算(Weather-Cloud)的启发式调度算法。调度算法对气象作业按照时间紧迫型、CPU紧迫型、内存紧迫型和硬盘空间紧迫型进行分类,计算资源综合紧迫指数,相应地赋予不同调度优先权限。与CMMS(Cloud Min-min Scheduling)、AFCFS(Adaptive First Come First Service)、Fair的调度算法对比表明,Weather-Cloud的调度算法不但减少了计算的等待时间,而且增加了完成的指令数量。

Based on the analysis of the characteristic of meteorological computing, we propose a cloud model for meteorology research and computing. At the same time, a resource schedule method-weather cloud is proposed for the model. In accordance with the time constraints, CPU limit, memory limit, disk space constraints, scheduling algorithm classify the meteorological jobs, computing integrated parameters of resources, corresponding scheduling priority given different jobs. Comparisons are executed to test the performance of our method and other methods. Simulations show that our method not only reduces the waiting time, but also enhance the finished instructions number.