K.Ganesan, K. Krishneswari
MapReduce is emerging as an important programming model for large-scale data-parallel applications such as web indexing, data mining, and scientific simulation. MapReduce runtimes like Hadoop tolerates only crash faults, not arbitrary or Byzantine faults. Byzantine fault tolerance algorithm in MapReduce typically requires 3f+1 servers to tolerate f Byzantine Servers, which involves considerable costs in hardware, software, and administration. By achieving arbitrary fault tolerance using proposed Byzantine Fault Tolerance (BFT) MapReduce algorithms, which improve previous algorithms in terms of several metrics. First, design a framework requires only 2f + 1 replicas, instead of the usual 3f + 1. Second, improve the performance of framework with help of non-speculative and speculative algorithms. An important aspect in terms of BFT MapReduce algorithm executes the job with acceptable cost for many critical applications.
Hadoop, MapReduce, Byzantine Faults