Simulated Annealing-Based Energy and Traffic-Aware Virtual Machine Placement in Fog Computing
Ibrahim Dhaini, Soha Rawas, Ali El‐Zaart · 2025
The fast expansion of the Internet of Things (IoT) is resulting in an unprecedented rise in data output, necessitating significant computer resources for processing and analysis. Fog computing (FC) complements cloud computing by bringing data processing closer to IoT devices, lowering network congestion, and improving system responsiveness. However, effective resource management, particularly virtual machine placement (VMP), remains a substantial difficulty in fog environments. This study looks at the power consumption of fog computing systems, specifically how VMP methods affect server and network energy usage. A simulated annealing-based VMP approach is suggested to optimize VM allocation across physical machines (PMs), reduce overall power consumption, and improve network traffic efficiency. A comparison of three VM allocation rules (LFF, MFF, and EnergyNetAware2-SA) reveals that EnergyNetAware2-SA consumes the least amount of energy across different workloads, outperforming previous techniques. It delivers up to 10% lower server power usage than LFF and 29% lower than MFF, while decreasing network power consumption by up to 20% and 33%, respectively. Furthermore, total power consumption is lowered by up to 14% versus LFF and 29% versus MFF.