Optimizing Virtual Network Function Migration in MEC through Graph-Based Resource Grouping

With the surge in mobile network traffic, the integration of Mobile Edge Computing and Network Function Virtualization has become a crucial technology to meet users' high-quality service demands. However, in a mobile edge computing environment, the migration of Virtual Network Functions has become an essential means of ensuring service quality, due to the constraints of limited resources and the high mobility of users. To address the issue of latency degradation caused by frequent user migrations, this paper proposes an optimization approach for VNF migration based on graph partitioning. First, users are grouped and resources are allocated based on their mobility differences. Then, the Adaptive Simulated Annealing (ASA) algorithm is applied to determine the optimal number of clusters that minimizes latency, and graph partitioning is performed based on this cluster number to ensure the minimal number of users migrating across clusters. Finally, migration algorithms are executed only when users migrate across clusters, effectively reducing unnecessary resource consumption and latency. Experimental results show that, compared to CLO-VNF, the proposed method reduces latency by 39.37%, demonstrating its effectiveness in minimizing end-to-end latency.