MaMED: ML-Assisted Minimum End-to-End Delay Routing in SDN-IoT Networks for IoT Monitoring

Incorporating software-defined networking (SDN) technology into Internet of Things (IoT) can improve network utilization. In this paper, we consider SDN-IoT networks, in which IoT devices use Zigbee protocol. A machine-learning- assisted minimum end-to-end delay (MaMED) routing scheme for IoT monitoring services is proposed; and this scheme requires no prior knowledge of input traffic. Under different SDN- IoT simulation scenarios, results show that our routing scheme outperforms the shortest path first (SPF) algorithm in terms of end-to-end delay (E2E) performances (e.g., average E2E delay and jitter), and gives more consistent delay performance when the input IoT traffic increases.