Energy-Efficient Virtual Machine Allocation and Migration Using a Dual-Threshold MBFD Strategy in Cloud Computing

Efficient Virtual Machine (VM) allocation and migration are critical in optimizing cloud resource management by ensuring load balancing and minimizing power consumption. Traditional VM placement strategies, such as Best Fit Decreasing (BFD) and Modified Best Fit Decreasing (MBFD), often fail to dynamically regulate migration, leading to excessive power consumption and performance degradation. This paper proposes a Conditional Dual-Threshold Policy, integrated into MBFD, to manage VM allocation and migration intelligently. The proposed approach is divided into two segments: the first segment focuses on VM allocation using a minimum cost policy, where VM placement decisions are based on pending buffers and execution costs to optimize initial resource allocation. The second segment applies a dual-threshold strategy within MBFD to migrate VMs from overutilized PMs while ensuring that target PMs do not exceed safe utilization levels. Experimental results demonstrate that the proposed approach outperforms both BFD and MBFD by achieving up to 22% lower power consumption compared to BFD and 14% compared to MBFD, while also optimizing workload balancing. These findings confirm that the Proposed Dual-Threshold MBFD Algorithm is a scalable and energy-efficient solution for dynamic cloud environments.