Optimizing Load Balancing and Minimizing Communication Latency in Edge Networks

Edge computing networks hold the potential to elevate service quality by minimizing communication latency. This is accomplished through the strategic placement of cloud services close to end users and employing dispersed servers located at the network's edge. Nevertheless, the deployment of edge servers should be closely aligned with the demand for computational resources and the distribution of user loads. As a result, strategies are required that optimally select the appropriate access nodes for hosting computing services and efficiently pair each remaining access node with the most suitable service node for service provisioning. In this paper, we conduct a simulation study on optimizing load balancing and communication latency in edge service networks. We propose an edge service node strategy composed of two schemes, namely the Service Node Selection (SNS) and Access Node Allocation (ANA) and compare it against Forward Greedy, a previously proposed heuristic, enhanced with Load Balance capabilities (FG-LB). We investigate the behavior of these two algorithms in a grid network topology, considering uniform workloads distribution. The results validate the efficiency of our proposed schemes in terms of latency and load distribution, in addition to the significant lower utilization of total processing capacity.