Accelerating Feedback Control for QoE Fairness in Adaptive Video Streaming Over ICN
Rei Nakagawa, Satoshi Ohzahata, Ryo Yamamoto · 2024
Today, information centric networking enables adaptive video streaming clients to further improve QoE by applying flexible content-based control. However, an adaptive bitrate algorithm makes a client occupy the bottleneck link at excessively high bitrate, reducing the QoE fairness to other clients sharing the bottleneck link. Then, we propose fairAccel, a method of accelerating bitrate-based feedback control for achieving QoE fairness. fairAccel assigns more bandwidth to clients selecting the lower bitrate while suppressing content requests from clients selecting the highest bitrate on the bottleneck link. In addition, to further improve QoE fairness, fairAccel exploits the symmetric routing of ICN content request / response and applies bidirectional feedback control to the content request / response path. Thus, fairAccel accelerates feedback control by mitigating router queues under control of suppressing content requests before excessive traffic is delivered to the response path. Through simulation experiments, fairAccel improves the average bitrate and further improves QoE fairness for representative ABR algorithms.