StreamGuard: Exploring a 5G Architecture for Efficient, Quality of Experience-Aware Video Conferencing

Researchers from Princeton University (Xuyang Cao, Oliver Michel, Kyle Jamieson) have published StreamGuard, a novel 5G architecture designed to dramatically improve video conferencing quality under constrained radio conditions. The system tackles a core problem: in crowded 5G networks, interactive traffic like video calls suffers because not all data packets are equally important to the user's experience. StreamGuard introduces a closed-loop control system with three key components: a monitor in the Radio Access Network (RAN) that uses deep packet inspection to infer QoE and network state, a controller that selects prioritization actions balancing QoE and fairness, and a marking module that steers subflows into appropriate priority queues. It also shapes application behavior through selective subflow dropping and probe-based rate control.

In real-world tests on a 5G testbed, StreamGuard achieved a superior QoE-fairness tradeoff compared to standard 5G and prior state-of-the-art approaches. The results show it can improve QoE by up to 70% at comparable background throughput, or preserve up to 2x higher background throughput at similar QoE levels. The 31-page paper, submitted to arXiv on April 29, 2026, includes 35 figures and falls under networking, multimedia, and image/video processing categories. This work represents a practical, implementable solution for the growing challenge of maintaining high-quality video conferencing in dense 5G environments.