MapReduce++: Ecient Processing of MapReduce Jobs in the Cloud ?

MapReduce is the most popular processing framework and it has gotten a big successful in the large scale applications on the clusters. In this paper, we proposed a new massive data processing framework MapReduce++. We focus on the massive MapReduce tasks’ ecient processing and hope to get a good response time for these tasks’ total time. To archive this goal, we designed an ecient processing algorithm of MapReduce tasks. We proposed a time cost model for the massive tasks’ processing. And our algorithm will consider all tasks’ time cost base this cost model of massive tasks processing, and nd a better tasks’ executing strategy and reduce all these tasks’ whole time cost. Simulation experiment showed that our algorithm will improve the eciency a lot.