An effective VM migration scheme for reducing resource fragments in cloud data centers

Abstract

Data center is an essential infrastructure in cloud computing which offers different types of resources for the services of cloud computing. An important issue is how to effectively use the resources in data centers. In order to improve resource utilization, virtualization techniques have been widely used in data centers. Several virtual machines (VMs) are concentrated on a server, and resource utilization of server can be promoted effectively. However, server utilization is inefficient in data centers. In this paper, we investigate how to reduce resource fragments and allocated resources in data centers. Hence, we propose a new scheme to estimate VM migration to effectively reduce resource fragments and allocated resources. The scheme improves the server utilization in data centers. Moreover, our proposed scheme can be applied to multiple resources. We use 5800 servers to proceed VM migrations in the simulation experiments. The simulation results show our scheme reduces 8% resource fragments and 3% allocated resources.