Adaptive Video Streaming with Scalable Video Coding using Multipath QUIC

The multi-path protocol improves end-to-end throughput and becomes one of the solutions to improve video streaming Quality of Experience (QoE). Scalable Video Coding (SVC) encodes a video segment into multi-layer, and different layers can be transmitted on heterogeneous paths. SVC has great potential to make use of the multipath protocol. Meanwhile, multipath QUIC (MPQUIC) provides multiplexing and instant handshake compared with multipath transmission control protocol (MPTCP). Thus MPQUIC has great potential in Dynamic Adaptive Streaming over HTTP (DASH) with SVC. However, mismatches between streams with different priorities and paths, network congestion caused by the suddenly increased data traffic are new challenges of MPQUIC for DASH-SVC. Adaptive Stream-scheduler Multipath QUIC (ASMQ) framework is proposed to improve the user's QoE in DASH-SVC, prioritizing the streams based on the DASH-SVC application information and evaluates the qualities of multipath. ASMQ schedule prioritized streams to the path with different qualities. In addition, the server-client feedback mechanism of ASMQ can adapt to network congestion conditions in real-time. Emulation with real network traces demonstrates that the ASMQ improves user's QoE compared to MPTCP and traditional MPQUIC.