Dynamic Resource Allocation Exploitation for Parallel Data Processing Efficiency in Cloud – A Review

Parallel data processing has become more and more reliable phenomenon due to the realization of cloud computing,especially using IaaS (Infrastructure as a Service) clouds. The cloud service providers such as IBM, Google, Microsoft andOracle have made provisions for parallel data processing in their cloud services. Nevertheless, the frameworks used as ofnow are static and homogenous in nature in a cluster environment. The problem with these frameworks is that the resourceallocation when large jobs are submitted is not efficient as they take more time for processing besides incurring more cost.In this paper we discuss the possibilities of parallel processing and its challenges. One of the IaaS products meant forparallel processing is presented in this paper. VMs are allocated to tasks dynamically for execution of jobs. With proposedframework we performed parallel job processing which involves Map Reduce, a new programming phenomenon. We alsocompare this with Hadoop.