Power Aware Scheduling of Virtual Machines Using Soft-Computing Techniques in Cloud Environment: A Framework for Energy-Efficient Resource Management

Abstract

With the increasing demand for cloud computing resources, efficient scheduling of Virtual Machines (VMs) has become a critical challenge. Traditional VM scheduling mechanisms often lead to high computational and memory overheads, resulting in increased energy consumption and compromised performance. This paper proposes a power-aware VM scheduling approach using soft-computing techniques, such as Genetic Algorithms (GAs), Fuzzy Logic, and Neural Networks. The proposed model leverages predictive resource utilization and intelligent VM placement strategies to optimize energy consumption while meeting Service Level Agreement (SLA) requirements. The model integrates the use of historical resource data, real-time workload prediction, and dynamic VM migration to reduce idle host energy consumption, prevent SLA violations, and improve overall cloud system performance. Experiments conducted using the CloudSim simulator show that the proposed approach effectively reduces power consumption, enhances resource utilization, and minimizes SLA violations compared to existing scheduling techniques. This research contributes to the advancement of power-aware scheduling in cloud data centers, paving the way for greener and more efficient cloud computing environments.