Increasing performance of parallel and distributed systems in high performance computing using weight based approach

High performance computing (HPC), large scale instruments and continuously increasing simulation tools are generating data at a huge rate that are difficult to be effectively managed and analyzed. Implementation of MapReduce model provides a way for processing huge volumes of data through the use of large number of commodity computers. MapReduce and Hadoop have been initially used for processing web data. But recently they have been used for processing more complex scientific applications. The proposed system helps to understand the impact of file system, network and programming modes on performance. The performance an application can obtain is largely work load dependent. Design of every MapReduce system has to include the Key factors like High speed, Quick Response, Accurate result. The proposed work is to improve the scheduling and management functionality of Parallel and Distributed Computing. The proposed technique Weight based Approach improves the performance by improving job scheduling strategy.