Dynamic Threshold-Based Resource Management for Fog Computing Environments

Fog computing extends cloud services to the network edge, thereby reducing latency and bandwidth usage for time-sensitive applications. However, the limited computational capacity, memory, and bandwidth of fog nodes present significant challenges for efficient resource management. This paper proposes a Dynamic Threshold-based Resource Management (DTRM) strategy that dynamically adjusts resource allocation thresholds to prioritize real-time (RT) requests while minimizing the redirection of non-real-time (NRT) requests to cloud servers. To further support balanced workload distribution, a Quota-Based Round-Robin (QRR) scheduler is introduced, ensuring fairness and low computational overhead across fog nodes. Extensive experimental evaluations demonstrate that DTRM significantly reduces the resource redirection count compared to Static Threshold (ST), Deficit Round Robin (DRR), and Best-Effort (BE) schemes. Moreover, DTRM improves load balancing and resource usage, offering a scalable and adaptive solution for dynamic and bursty request patterns in fog computing environments. These results highlight the potential of DTRM to enhance the overall performance, responsiveness, and efficiency of fog computing, particularly in the context of modern IoT applications.