应用于云中MapReduce计算的动态资源调度算法
The Dynamic Resource Scheduler for MapReduce in the Cloud
-
摘要: 越来越多的MapReduce计算由运行在云中虚拟机组成的虚拟集群完成,为增强虚拟MapReduce集群计算中任务的数据本地性,充分利用云计算的资源可动态配置的优势,提出了一种基于队列的动态资源调度算法,并模拟了云计算平台进行测试,实验结果表明动态资源调度算法提高了虚拟MapReduce集群的计算性能.Abstract: It has been increasingly adopted that MapReduce jobs runs on a set of virtual machines (VMs) in the cloud.To enhance the data locality of the tasks during the job,leverages the advantage of the cloud's dynamic resource reconfiguration,this paper proposed a Queue-based dynamic resource schedule algorithm.Cloud platform had been simulated for the evaluation and the result shows that the scheduler efficiently improved the virtual clusters' performance.