Multi-Objective Algorithms for Virtual Machine Selection and Placement in Cloud Data Center

Abstract

Dynamic Virtual Machine (DVM) consolidation is one of the operational techniques to reduce energy consumption by grouping Virtual Machines (VM) in a limited number of hosts. However, the mighty Virtual Machine consolidation may incite the Service Level Agreements violation (SLAV). In this study, we proposed (a) Multi-objective Virtual Machine selection (MOVMS) algorithm to determine a migratable Virtual Machine from the overloaded host and (b) Multi-objective Modified Best Fit Decreasing (MOMBFD) algorithm for virtual machine placement. We have implemented it in cloudsim and the result shows that the proposed approaches appear to significantly reduce energy consumption by 31.36%, Service Level Agreements violation(SLAV) by 4.95%, and decrease the number of migrated Virtual Machine by 28.49 % compared to some previous studies which used Maximum Correlation (MC) algorithm, Minimum Migration Time (MMT) algorithm, Minimum Utilization (MU) algorithm and Random Selection (RS) algorithm for Virtual Machine selection and Power-Aware Best Fit Decreasing (PABFD) algorithm for Virtual Machine placement.