Bi-Objective Optimization of Bandwidth Resources and Energy Consumption for Efficient Virtual Machine Placement in Cloud Computing

Abstract

Cloud computing has revolutionized IT organizations through its on-demand computing model, with Infrastructure as a Service (IaaS) being one of the leading services. Cloud service providers maintain a large number of physical devices, including servers and networking equipment, to meet user demands. However, these devices consume a significant amount of energy and require sufficient bandwidth to handle data traffic during virtual machine placement. Balancing energy consumption and bandwidth resource utilization is crucial for CSP, and user experience. To address this trade-off, this thesis introduced BOHGOA integrated with ACO for efficient virtual machine placement in cloud computing. It aims to optimize the placement process by considering both bandwidth resource utilization, and energy consumption simultaneously. The proposed approach generates a set of Pareto-optimal solutions, providing a systematic approach of non-dominated solution selection for the bandwidth resource-energy trade-off. The performance of BOHGOA was evaluated using the CloudSim toolkit, comparing it with GA, ACO, and FFD algorithms. The results demonstrated the effectiveness of BOHGOA, showing significant reductions in bandwidth resource consumption by 54.14%, 32.11%, and 57.47% for 240 VM placement and 38.12%, 22.76%, and 47.05% for 500 VM placement when compared with GA, ACO, and FFD, respectively. Additionally, the proposed algorithm achieved notable reductions in physical machine energy consumption by 37.70%, 34.01%, and 40.14% for 240 VM placement, and 27.50%, 22.28%, and 30.52% for 500 VM placement when compared with GA, ACO, and FFD, respectively. These findings highlight the effectiveness and potential benefits of employing BOHGOA as a solution for optimizing virtual machine placement in cloud data centers when compared with these mentioned three algorithms.