Abstract

Cloud computing play an important role in data intensive application since it provide a consistent performance over time and it provide scalability and good fault tolerant mechanism Hadoop provide a scalable data intensive map reduce architecture Hadoop map task are executed on large cluster and consumes lot of energy and resources Executing these tasks requires lot of resource and energy which are expensive so minimizing the cost and resource is critical for a map reduce application So here in this paper we propose a new novel efficient cloud structure algorithm for data processing or computation on azure cloud Here we propose an efficient BSP based dynamic scheduling algorithm for iterative MapReduce for data intensive application on Microsoft azure cloud platform Our framework can be used on different domain application such as data analysis medical research dataminining etc Here we analyze the performance of our system by using a co-located cashing on the worker role and how it is improving the performance of data intensive application over Hadoop map reduce data intrinsic application The experimental result shows that our proposed framework properly utilizes cloud infrastructure service management overheads bandwith bottleneck and it is high scalable fault tolerant and efficient