AutoPlex: inter-session multiplexing congestion control for large-scale live video services

Bo Wu, Tong Li, Cheng Luo, Changkui Ouyang, Xinle Du, Fuyu Wang

Abstract: The increasingly obvious advances in live video services introduce the urgent need for enhancing network transmission performance, especially by designing an efficient congestion control (CC) scheme. Unfortunately, the previous rule-based CC methods cannot adapt well to various network conditions and statuses while machine-learning-powered CC paradigms always suffer from non-trivial system overhead and unstable effects. In this paper, we first conduct a large-scale network measurement for 800+ million live video streams, and find that QoS metrics of better-performed sessions show similarity in the same user group. We then propose AutoPlex, an inter-session multiplexing CC framework that makes full use of this similarity and automatically adjusts CC parameters (i.e., pacing rate and congestion window size). AutoPlex supports user-defined policies that can act as standards to learn QoS features of better-performed sessions. We implement the proposed AutoPlex prototype based on QUIC protocol and BBR algorithm, and conduct experiments in the real live CDN proxy. The experimental results demonstrate the potentials of AutoPlex for the transmission optimization of live video applications, in which the average (or 90th-percentile) retransmission ratio can be reduced by 24% ~ 27% (or 32% ~ 40%) while the average value of goodput/rtt is promoted by 14% ~ 32%.