Performance Improvement Issues and Approaches for Hadoop

Abstract

Nowadays Hadoop is a go-to framework for Big Data Analytics. In current scenario data is growing exponentially and Hadoop is the defacto solution for this growth. Although Hadoop is popular for its  high-performance computing in data-intensive applications, increasing evidence has shown that performance of data-intensive applications can be severely limited by many factors like hardware structure of nodes, Algorithmic strategies used, architectural decisions,  and also whether it is running on physical server or virtual. This paper incorporates performance improvement in terms of movement of data between nodes, proper utilization of resources like CPU and I/O. The attempt of machine learning for improved resource utilization is also included which takes us to the new era of performance improvement.