Energy aware resource management and job scheduling in cloud datacenter

Abstract

A cloud system uses virtualization technology to provide cloud resources (e.g. CPU, memory) to users in form of virtual machines. Job requests are assigned on these VMs for execution. Efficient job assignment on VMs will reduce the number of hosts used. Hence, it is essential to achieve energy optimization in cloud computing environments. Therefore, in this paper, a job scheduling mechanism is proposed to assign job to a VM of the existing active hosts itself by considering job classification and preemption. So that minimizing the number of host used in allocation intern reduces the energy consumption in the Cloud datacenter. In our proposed job scheduling algorithm, categorizing the job in to three different types and assigned based on preemption policy with the earliest available time of the resource (VM) which is attached to a host. Thereby, we reduce the energy consumption by making less number of hosts in the active state and increase the utilization of active host. Finally, we conduct simulations using CloudSim and compare our algorithm with other existing methods. Significant energy savings can be obtained depending on system loads. Energy saving is about 2% to 46% with respect to the non-energy aware algorithm, 1% to 7% than the energy aware algorithms.