A Novel Parallel Computation Model with Efficient Local Memory Management for Data-Intensive Applications

The provisioning of high-performance computing infrastructure through cloud environments enables data intensive processing to be a viable solution. In this paper, we introduce a novel parallel computation model similar to MapReduce framework. The proposed parallelized model incorporates a parallel execution strategy in worker nodes to decrease execution response times in cloud environments. The parallelized model adopts efficient local memory management techniques in the worker nodes to reduce memory transfer overheads. For evaluation, we compared the proposed framework with the state of art Hadoop MapReduce framework. From experiments on benchmark datasets, it turns out that the parallelized model reduces the execution times by about 45.86%. Those experimental results indicate the efficiency and the scalability of proposed framework on cloud environments.