DASH+: Download Multiple Video Segments with Stream Multiplexing of QUIC
Changjiang Cui, Yifei Lu, Shuren Li, Jingqi Li, Zeqi Ruan · 2022
The emerging cloud computing and big data technologies provide enlightenment to intelligent processing for large-scale video data. MPEG-DASH (DASH) is becoming the unified adaptive bitrate (ABR) streaming open-source standard in video streaming. However, since DASH runs over TCP, it can lead to Head-of-Line (HOL) blocking problem that seriously degrades network performance. Moreover, we find that the advantage of multiplex transmission for Quick UDP Internet Connections protocol (QUIC) can tackle HOL completely. But the traditional DASH algorithm can not support multiplex transmission because it can only transmit one segment every time. Therefore, in this paper, we propose DASH+, a modified DASH algorithm to support the multiplex transmission of QUIC. DASH+ allows clients to request multiple video segments on separate streams using the same QUIC connection. When the playback buffer occupancy is sufficient, the segments from the server are adaptively transmitted through multiple streams according to the corresponding deadline. We conduct the simulation under ns3 and compare DASH+ with the traditional DASH algorithm under different link delays. The results show that under the throughput-based ABR algorithm, DASH+ over QUIC performs higher throughput and better video quality than the others. And the higher the link delay, the more pronounced the advantage.