Clustering-based Redundancy Minimization for Edge Computing in Future Core Networks

The increasing demand for new and heterogeneous services generates redundant signalling, leading to communication overheads and congestion in the network’s core. We propose a novel AI-enabled edge architecture to minimize signalling redundancy. We deploy a cluster-based signalling and admission control framework to maximize the efficiency of link (or bandwidth resources) between the edge and core networks. We minimize redundant signalling using two classical unsupervised machine learning algorithms (K-mean and Ranking-based clustering). Our results show that the proposed framework provides substantial latency reduction while maximizing resource utilization. The proposed approach is 35% superior in reducing redundant signalling compared to recent work.