Performance comparison of Hadoop Clusters configured on virtual machines and as a cloud service

Abstract

Cloud computing is an emerging trend for online computing and resource management. OpenStack is one of the most widely used open source platform for building public and private clouds. Whereas, Apache Hadoop is an open source framework use to process large data sets spread across clusters of computers. Currently both OpenStack and Apache Hadoop are sharing the major consideration in research and open source communities. This paper focuses on integration of Hadoop Cluster on OpenStack cloud as it's one of the services and then highlights the performance comparison of Hadoop cluster implemented with that of Hadoop cluster configured separately. The results concluded show the successful implementation of Hadoop as a cloud service and its performance enhancement in comparison to native virtual Hadoop cluster.