Modified Energy-Aware Rolling Horizon algorithm for scheduling of cloudlets

Abstract

Energy conservation through efficient energy utilization is important to minimize energy cost and thereby reduces energy demand. The existing systems support either system schedulability or energy conservation. If it priorities energy conversation, then there is a possibility of rejecting a task due to poor schedulability. In the existing system, some tasks are able to complete using minimum virtual machines. While doing so, other virtual machines will be in ideal mode; hence substantial energy is wasted for idle virtual machines while performing these tasks. Our proposed system named Modified Energy Aware Rolling Horizon ie. MEARH which schedules realtime tasks in the rolling horizon. MEARH algorithm mainly focuses on user's timing requirements by completing as many tasks as possible within the deadline without rejecting a single task by optimizing available resources. Energy conservation shall be achieved by consolidating ideal virtual machines and shall be switched to sleeper mode. Thus it improves the performance, scalability and availability of resources. MEARH's optimization policy integrates other energy aware algorithms. Comparison results with other scheduling algorithms show that MEARH significantly improves the scheduling quality of real-time tasks without rejecting a single task.