Improving mapreduce scheduling algorithm using prediction-based application model in the cloud

Cloud computing has been a new paradigm for providing computing resources and services in many areas including academy and industry. The Cloud providers oeffrer various types of service, such as Infrastructure as a Service (IaaS), Platform as a Service (PaaS), and Software as a Service (SaaS). One of important issues in Cloud computing environments is efficient resource or service management of huge resource in Cloud computing resource farm. In this paper, we provide an efficient scheduling algorithm of Map Reduce framework which is an emerging application model in Cloud computing. The proposed scheduling algorithm schedules map and reduce works using intermediate key production information of map works in heterogeneous computing environments. Throughout simulation results, we show that the proposed one improves scheduling performance compared with current Map Reduce scheduling algorithms.