An Novel Parameter Tuning Alogrithm for MapReduce Model

Abstract

The performance optimization of MapReduce clusters has received significant attention in recent years, because it plays an important role on MapReduce clusters. Parameter tuning is one important way to optimize the performance of MapReduce clusters. Traditional parameter tuning performs poorly in automatic configuration of the parameters. To address the problem, an efficient parameter tuning algorithm for MapReduce clusters is proposed in this paper. In this paper, a detail analysis about the execution process of jobs on MapReduce clusters is described. Moreover, the objective function about the parameters is introduced by the least square method. In order to solve the objective function, particle swarm optimization is introduced to find the optimal solution of the parameters. Experimental results prove that the proposed parameter tuning algorithm is able to perform well in terms of execution time for jobs in MapReduce clusters.