A Virtual Network Function Migration Mechanism in Software-Defined Networks Using the Louvain Algorithm
Simin Mehrasa, Vahid Ayatollahitafti · Artificial Intelligence Tools in Software and Data Engineering (AITSDE) · 2025
This paper presents a comprehensive framework for optimizing Virtual Network Function (VNF) deployment and migration in multi-domain NFV environments. The proposed algorithm, based on an adapted version of the Louvain method, jointly optimizes VNF placement and migration while simultaneously considering latency, reliability, throughput, and resource efficiency. The framework employs a three-layer hierarchical control architecture comprising edge, domain, and root controllers: the modified Louvain algorithm identifies optimal communities based on latency criteria at the edge layer, domain controllers are deployed considering load balancing and strategic positioning, and the root controller is placed at the node with the highest betweenness centrality. Simulation results on a real-world topology demonstrate that the proposed approach reduces end-to-end delay, increases service reliability, improves resource efficiency, and lowers migration costs compared to conventional methods.