A Novel Approach to Improving Resource Utilization for IaaS

Abstract

IaaS is one of the three major cloud computing service models. Due to the dynamic arrival and departure of VM instances in IaaS, computing resource fragments cannot be avoided by traditional placement algorithms such as FFD and BF. In this paper, therefore, we propose a novel VM replacement algorithm called MCARD (Migration Cost-Aware Resource Defragmenter) to solve the problem that how to maximize the overall resource utilization for IaaS under the given migration cost constraint. In addition, based on the main idea of MCARD, we also present a new VM placement algorithm named MPlace, which is able to efficiently increase the resource utilization through properly using the migration constraint. The experimental and simulation results show that MCARD can increase the resource utilization by up to 7.27% while MPlace is able to improve the utilization by 2.83% on average compared with traditional VM placement algorithms.