Adaptive Node Management in Edge Networks Under Time-Varying Workloads

The emergence of edge computing architecture offers a promising prospect for augmenting service quality and user experience across a wide range of applications, by mitigating communication latency. This approach adopts the deployment of computational resources closer to the data source, at the edge, enabling faster response times and improved performance, critical for applications demanding real-time data processing. However, operational service nodes that are either underutilized or overutilized deteriorate system performance and impose a significant burden on energy consumption. Thus, it is important that the system adapt quickly to time-varying workloads and reconfigure service nodes accordingly. In this research, we simulate and study an edge network under time-varying workloads and propose a procedure to monitor service nodes utilization and increase or decrease the computing capacity based on heuristic schemes that integrate both workload balance and service proximity for selecting appropriate service nodes and assigning access nodes to service nodes. Simulation results demonstrate that our proposed methodology and schemes manage to keep system operation at the desired levels, while evenly distributing workload across service nodes and keeping communication latency and service migration low.