An Enhanced and Efficient Approach for Improving the Performance of HPC Environment Using Map-Reduce With MARIANE

Hadoop is the main open source tool which backs the acknowledgment of the Big data upheaval and depends on Googles MapReduce spearheading mission in the area of very much substantial measure of information stocking space and handling. As of late, functions and advancements tending to "Big Data" confrontations have turned out to be colossally critical in numerous ranges of business and logical registering. The Map-Reduce model has empowered an expansive arrangement of uses, has prodded the improvement of open-source programming advances such as Hadoop, and is making a whirlwind of enthusiasm for the scholarly in addition to researchers from industry around enhancements and optional ways to deal with taking care of the requests of breaking down extensive complex information sets. The HPC people group is constantly vigilant for approaches to influence more extensive product speculations to distinguish mass-market advances which could be viably connected to take care of issues particular to high-performance computing. For example there are a few cases where outing of the rush of mass-business sector innovations has fundamentally affected HPC. Utilizing bigger business sector advances and the related economy-of-scale is being a compelling method for lessening prices and drawing in a bigger group. HPC, much competent devices, and regularly expanding reenactment scales are creating information at a pace that surpasses our capacity to viably oversee, break down, and distribute it. Execution of MapReduce pattern gives a method for extensive information volumes that could be consistently handled through utilization of vast item PCs. Hadoop and MapReduce have generally been utilized for web information handling and just as of late been utilized for exploratory applications. Our recommended framework comprehends the effect of document framework, programming modes on execution and system. Acquirement of an application's execution is to a great extent work load subordinate. Configuration of each MapReduce frameworks needs to incorporate the Key components such as, precise output, Quick Response and High speed. Enhancement of the Fault resilience of Map Reduce Implementation Adapted for High Performance Computing Environments in Distributed and Parallel Computing is considered to be the primary objective of this work. (C) 2017 Elsevier Ltd. All rights reserved.