Introduction

The Media over QUIC (MoQ) initiative, led by the Internet Engineering Task Force (IETF), aims to revolutionize live streaming by leveraging the QUIC protocol to achieve low latency and scalable media transmission. Traditional live streaming platforms like Twitch and YouTube use protocols such as RTMP for media ingestion and adaptive streaming over HTTP for distribution, which are effective for scaling but often result in high latencies. Conversely, real-time protocols like RTP provide low latency but are challenging to scale. MoQ seeks to develop a unified protocol stack that addresses both issues by utilizing QUIC's advanced features like improved congestion control and elimination of head-of-line blocking.

This thesis aims to implement a prototype live-streaming system based on the MoQ framework, allowing media to be streamed through QUIC from a server to clients with low latency and high scalability. The system will be compared to traditional streaming architectures to demonstrate its advantages in reducing latency and improving performance. This project highlights the potential of MoQ to enhance live streaming experiences, setting a new standard for interactive media applications.

Background

Current adaptive streaming methods, such as Dynamic Adaptive Streaming over HTTP (DASH) and HTTP Live Streaming (HLS), are optimized for HTTP/1.1 and HTTP/2, both of which rely on TCP [1]. However, when these methods run over HTTP/3 (H3), which uses QUIC as its transport layer, they do not perform significantly better [2, 3]. Besides, TCP suffers from Head of Line blocking and is not encrypted out of the box. Studies have indicated that QUIC can outperform TCP in congested environments, but to fully leverage QUIC's features, a custom application-layer protocol is necessary [4], and Media over QUIC Transport (MOQT) fills the gap. One recent research study also shows that MoQ yields much better latency(37ms vs. 2.993s) while only consuming 4.38% more bandwidth under the same bitrate configuration (3000 Kbps) in a testbed that uses traffic control for network emulation in a local area network [5].

We will use React combined with WebTransport on the frontend and quic-go for the thesis project to build the backend server. React will facilitate the development of a responsive and dynamic user interface. WebTransport is part of the modern web stack; it exposes QUIC-like APIs to web applications if HTTP/3 is supported or it falls back to HTTP/2; it inherits the characteristics of QUIC's performance, security, and reliability features while providing good support for web applications. The quic-go library will be used to facilitate building the backend server, offering a robust and efficient implementation of the QUIC protocol. This combination will enable us to build a low-latency live-streaming system that utilizes QUIC's advanced features.

Problem Statement and Research Questions

Traditional live streaming architectures, such as those used by platforms like Twitch and YouTube, suffer from high latency despite their scalability, negatively impacting interactive applications. The Media over QUIC (MoQ) initiative aims to address this issue by leveraging QUIC's advanced features for low-latency and scalable media streaming. This thesis seeks to design and implement a prototype live-streaming system using MoQ, comparing its performance to traditional architectures to demonstrate its benefits in latency and scalability. Potential research questions come to mind:

• How can the client and server perform rate adaptation?
  • Rate adaptation is crucial for maintaining video quality and smooth playback in fluctuating network conditions. Both the client and server can perform rate adaptation to ensure optimal streaming performance. While the server can perform rate adaptation through advanced congestion control mechanisms provided by QUIC, it might increase latency, impact performance, and result in a degraded user experience.
  • From the other perspective, client rate adaptation can be achieved by monitoring buffer levels and network throughput to make informed decisions about which bitrate to request next. However, this will increase the client's complexity and result in potential inconsistency issues that different clients might experience inconsistent quality, which requires more in-depth research.
• How does adaptive bitrate streaming work for the client and server?
  • Adaptive Bitrate Streaming (ABR) allows the streaming system to dynamically adjust the quality of the video in real time based on network conditions. The server can encode the video at multiple bitrates and provide a manifest file that lists all available versions, or alternatively, the server can encode the video on demands of the client's choice.
• How can media be cached efficiently?
  • Efficient media caching is essential for reducing latency and ensuring smooth playback in a streaming system. Caching media can be done either at the edge server or end-user device or caching on both sides. Different media or frame types might have distinct cache persistence policies, such as I-frames (Intra-coded frames) can often be cached for longer durations compared to P-frames (Predictive-coded frames) due to their independence and critical role in video decoding.

Summary

This thesis aims to develop a low-latency live-streaming system using Media over QUIC (MoQ), leveraging its advanced features, such as low latency and high scalability, to improve the streaming experience compared to traditional architectures. The project involves implementing a prototype system that aims for low latency while also enabling efficient scaling by supporting adaptive bitrate streaming (ABR) for real-time adjustment of video quality based on network conditions or user demands. The system will utilize client-side rate adaptation to optimize the playback experience and dynamically switch between different bitrates as network conditions fluctuate. Efficient media caching strategies, such as segment-based or frame-level caching, will be employed to minimize latency and ensure smooth playback. While a common application layer protocol that functions as relays in CDNs realizes the scalability of our system.

Potential Challenges:

1. Protocol Adoption: Implementing MoQ involves designing and integrating a relatively new protocol stack into existing streaming infrastructures, which may require overcoming compatibility and integration challenges such as embedding WebTransport into the frontend in React and integrating Media over QUIC protocol into streaming.
2. Real-time Adaptation: Ensuring seamless client-side rate adaptation poses technical challenges, such as accurately predicting network conditions and minimizing switching disruptions.
3. Scalability: Managing scalability while maintaining low latency and high-quality streaming across a large user base requires robust server-side infrastructure and effective load-balancing strategies.

Addressing these challenges will be critical to realizing the potential benefits of MoQ for live-streaming applications, ultimately enhancing user experience through reduced latency and improved stream quality.

Reference

[1] A. Bentaleb, B. Taani, A. C. Begen, C. Timmerer, and R. Zimmer- mann. A survey on bitrate adaptation schemes for streaming media over HTTP. IEEE Communications Surveys & Tutorials, 21(1):562–585, 2019 (DOI:10.1109/COMST.2018.2862938).