An SDN-Enabled Framework for a Load-Balanced and QoS-Aware Internet of Underwater Things
Yaliang Shi, Qiuling Yang, Xiwen Huang, Deshun Li, Xiangdang Huang
Abstract
The massive demand for marine exploitation has promoted the thriving Internet of Underwater Things (IoUT). The volume, velocity, and variety (3V) of data produced by sensors, hydrophones, and cameras in IoUT are enormous, which challenges the network in achieving load balancing and Quality-of-Service (QoS) provisioning. This article adopts the “SDN+AI” paradigm to realize a load-balanced and QoS-aware software-defined IoUT from a framework design. We first introduce SDN technology to separate the data plane from the control plane to enhance the network’s scalability and flexibility. Then, a multicontroller load-balancing strategy based on switch migration called CASM is proposed to improve the network’s performance further. With the global view provided by SDN controllers, we proposed a QoS-aware adaptive routing protocol (SQAR) based on reinforcement learning, which can intelligently select route paths to satisfy the QoS requirements of multiple IoUT services. The results show that CASM achieves an efficient load balance while shortening the response time and average control path latency of the switch migration process, which significantly benefits our routing protocol. SQAR outperforms the existing QoS-aware routing protocols regarding QoS satisfaction probability, energy consumption, and convergence rate. Overall, our framework maintains a QoS violation rate below 5% and a load-balancing rate above 90% in a timely manner.