Maximum Satisfaction Scheduling algorithm Based on Hadoop Architecture
Kuan-ting Chen, Jianhua Huang, Yi Jin, Xi He · 2016
Based on the MapReduce job scheduling technology for design reference, this paper has put forward the maximum satisfaction scheduling algorithm of Hadoop to effectively solve the scheduling problems in MapReduce.The algorithm has tried to modify the original algorithm of Hadoop, configure a satisfaction score for each submitted job, and obtain the maximum satisfaction score of the job under the same Hadoop system environment of hardware and software.Compared with the own scheduling algorithm of Hadoop-fair share scheduling, the experiment has eventually drawn the conclusion that the maximum satisfaction scheduling algorithm can get the outcomes that customers want, reduce a certain degree of scheduling time, and enhance the system throughput.