A Multi-objective Virtual Machine Placement Optimization in Sustainable Cloud Environment

Virtual Machine Placement involves the selection of the optimal physical machine for deploying a requested virtual machine within extensive cloud data centers. VM placement must be performed strategically by considering different factors of the available resources for optimal exploitation of them. Numerous methods have been devised to address this issue. Nevertheless, existing solutions only account for a restricted set of resource types, leading to an uneven workload distribution that triggers the activation of unnecessary PMs within the data center. This work introduces a Non-dominated sorting genetic algorithm II that integrates various resource-constraint metrics to determine the optimal PMs for deploying VMs in a cloud environment. This algorithm maximizes resource utilization while minimizing the data center’s energy consumption and carbon footprints. The algorithm’s performance assessment is conducted using the Google Cluster Data set, and the outcomes are compared with established methodologies. The results showcase a substantial reduction in energy consumption, carbon footprints, and number of active PMs by 19.29%, 47.50%, and 79.81%, respectively. Furthermore, there is a noteworthy enhancement in resource utilization, reaching up to 67.28%.