A Traffic Prioritization Framework For Smart Home IoT Networks using Programmable Data Planes

The rapid growth of IoT provides a wide range of applicationsx, such as smart homes, healthcare, industry, etc. However, these applications generate various data that must be processed dynamically to ensure the desired QoS. Fortunately, recent advancements in data plane programmability provide some unique opportunities to dynamically process the loT traffic flow in a fine-grained manner. However, most research lacks of provide a solution to prioritize and schedule loT traffic flows in programmable switches. This paper proposes a dynamic prioritization technique for loT traffic flows by analyzing header field values granularly. Further, a weighted fair queue scheduling algorithm is designed in the programmable switch, which updates queue weights dynamically to reduce the delay for delay-sensitive loT applications. The experimental results show that the proposed solution prioritizes loT traffic flows 99.6% correctly. Additionally, the solution has achieved a throughput of 410.377Mbps for high-priority queues, and high-priority flows experienced a minimum delay of 1.84ms on average.