Dynamic Service Chain Migration for Network Functions Virtualization

Virtual network function supports the scalable control to the software-based network functions while leading to the challenge of flexible service chains migration for the load balance, processing performance or QoS. In this paper, we formulated the service chains migration (SCM) problem as an integer linear program that maximizes the utility of the network and subjects to the resource and performance constraints. We approximated the objective function of SCM with the Log-Sum-Exp function, and proposed a Markov chain-based distributed migration algorithm. Simulation results prove that the proposed algorithm can realize effective and cost-efficient service chain migration.