Preemptive flow management in future SDNized wireless networks

The emerging Software-Defined Networking (SDN) concept aims at simplifying network management and control. SDN is envisioned as one of the core technologies for future wireless networks. However, such networks have inherent challenges, in particular mobility management (MM) and network selection, which require careful design in a centrally controlled SDNized network. In this paper, we propose a proactive flow management solution based on an algorithm that learns users' mobility behavior and proactively processes the handover to reduce the connection delays. The proposed approach was evaluated on a testbed that employs OpenDaylight controller, Mininet emulator, and a custom-built control interface. The results of experiments on the testbed verify the convergence and scalability of the learning algorithm, and showcase the potential of proactive flow setup by the SDN controller in improving the handover delays.