An Approach for Service Function Chain Reconfiguration in Network Function Virtualization Architectures

Network function virtualization (NFV) technology continues gaining attention as a paradigm shift, and telecommunication services can be flexibly deployed and managed, resulting in lower CAPEX and OPEX for the Internet service providers (ISPs). Any telecommunication service is represented by a so-called service function chain (SFC), which can be achieved by a set of virtual network functions (VNFs) to be executed on virtual machines (VMs) in a strictly order according to SLAs. The NFV-enabled SFCs applied in the future 5G networks gradually emerges a challenging problem, particularly more and more users are trying to access their telecommunication services whenever and wherever, SFCs are needed to be dynamically and adaptively reconfigured, thus ensuring lower resource consumption and higher revenues for ISPs. This problem is called service function chain dynamic reconfiguration problem (SFC-DRP), and an integer linear programming (ILP) formulation with implementation in CPLEX is provided to exactly solve this problem. To reduce the algorithm time complexity, a novel heuristic solution is provided that can balance the tradeoff between service provider's revenue and reconfiguration cost. The obtained simulation results show that our proposal can approximate the performance of the ILP within a polynomial time, and outperform the existing benchmark in terms of reconfiguration overhead and the revenue from service provisioning respectively.