Dynamic Packet Flow Allocation Method Considering Reconfiguration Overhead with Intermediate Solutions in Software- Defined Networks

Software-defined networking (SDN) is a network virtualization technology that realizes flexible network control by software operation using an SDN controller. There is an approach to formulate a packet flow allocation problem as a mathematical optimization problem and solve it. Intermediate solutions output from a mathematical optimization solver can be used to allocate packet flow routes before the optimal solution is obtained. At the time of a network reconfiguration, the overhead of the packet flow allocation occurs. Therefore, we need to select intermediate solutions that improve network resource utilization while suppressing overhead. This paper proposes a packet flow allocation method that applies intermediate solutions considering reconfiguration overhead in the network environment where the traffic amount changes dynamically. We devise a policy that takes into account both overhead and network congestion ratio, which is used in the proposed method. The proposed method is compared to the benchmark policies. The benchmark policies include one that does not use intermediate solutions and one that applies intermediate solutions each time without considering any overhead. Numerical evaluations show that the proposed method reduces the time-averaged network congestion ratio while reducing overhead compared to the benchmark policies.