Accelerating Iterative Big Data Computing Through MPI
Fan Liang  Xiaoyi Lu
In this paper, we first analyze the overhead of shuffle operation in Hadoop and Spark when running PageRank workload, and then propose an e