Best Fit Decreasing Algorithm for Virtual Machine Placement Modeled as a Bin Packing Problem

This paper presents a novel Best Fit Decreasing (BFD) algorithm for virtual machine (VM) placement, formulated as a bin packing problem in cloud data centers. The algorithm focuses on minimizing resource wastage by strategically allocating VMs to physical servers. It operates in two stages: first, sorting the VMs in descending order of resource requirements, and second, searching for the best-fit server based on available resources. Extensive simulations using real-world workload traces demonstrate that the BFD algorithm outperforms popular placement strategies, reducing the number of servers required and achieving higher packing efficiency. Comparative analysis against First Fit Decreasing (FFD) and Best Fit (BF) algorithms reveals consistent superiority in terms of resource utilization and packing efficiency. The proposed BFD algorithm offers an effective solution for VM placement, improving resource utilization and system efficiency in cloud data centers. Its findings can guide the development of practical VM placement algorithms, leading to enhanced resource management and cost savings in cloud computing environments.