Does QUIC Suit Well With Modern Adaptive Bitrate Streaming Techniques?

摘要

This letter provides a thorough analysis of various recent adaptive bitrate streaming (ABR) techniques over Google's QUIC protocol. More specifically, we focus on the recent developments of the ABR techniques using control theoretic and reinforcement learning based approaches, and compare their performances over HTTP/TCP and HTTP/QUIC in terms of a wide range of video quality of experience (QoE) parameters. We observe that these ABR techniques are more compatible with HTTP/TCP compared to HTTP/QUIC, and the features of QUIC demands for new ABR technique which can balance the three fundamental QoE parameters - bitrate, smoothness and re-buffering during a video playback.