Optimizing Cloud Resource Utilization with ANN-Based VM Placement and Prediction

Cloud computing practitioners and the advancement of next-generation data centers face substantial challenges in terms of energy-related costs and environmental sustainability. To optimize this virtual machine (VM) placement is a better technique for minimizing energy while maximizing resource utilization. VM placement techniques often require knowledge of both current and future energy consumption, making it challenging to accurately predict the future demand of cloud applications. A VM placement strategy (MBFD) was employed, and its outcomes were utilized to train the proposed prediction model. The proposed prediction model considers both power consumption and CPU utilization of the physical machines (PMs). The results of the prediction model indicate a 10.96% minimized power consumption and a 6.9% improvement in service level agreement (SLA) violation.