SPV: Formal Verification of Stateful Parallel SFC Correctness

With the evolution of modern networks, network operators are able to flexibly deploy Service Function Chains (SFCs) in accordance with network policy requirements to achieve desired purposes. At the same time, to reduce the latency of processing packets through SFCs, an increasing number of network operators are designing SFCs to operate in parallel. In addition, many service functions are designed to be stateful, making the processing behavior of packets state-dependent, but the state is complex and arbitrary, which introduces greater complexity and uncertainty and makes the verification of SFCs behavior more difficult. To verify the correctness of stateful parallel SFCs, a formal verification method named SPV is proposed in this paper. SPV first defines the correctness attribute of stateful parallel SFCs, and then uses Coloured Petri Nets (CPN) for offline modeling of stateful parallel SFCs. This model describes the topological structure and behavior of the stateful parallel SFCs and enhances the expressiveness of the CPN model with Standard ML Language (ML) functions. In experiments, by analyzing the reachability graph and state space of the stateful parallel SFC's CPN model, the correctness attribute of the stateful parallel SFCs are verified, completing the correctness verification of stateful parallel SFCs. Moreover, these results also demonstrate the potential of CPN-based modeling for the verification of SFCs in stateful networks.