一种异构环境下的 Hadoop 调度算法

摘要

As an open-source implementation of MapReduce, the performance of Hadoop depends closely on its task scheduler. We design a new Hadoop scheduling algorithm for heterogeneous environments, which estimates the time of task completed more accurately, hasmorepertinenceofload,reducestaskresponsetimeandimprovestheperformanceofMapReduce.%　　作为MapReduce的开源实现,Hadoop的性能依赖于其任务调度机制.对Hadoop的推测执行算法进行研究,并设计了一种更具公平性、更好系统处理能力和更适合于异构环境的改进的Hadoop调度算法SALS.改进的Hadoop算法对调度任务的判定更准确,对负载的针对性更强,可减少系统对全部任务的响应时间,提高MapReduce的性能