Enhancing Real-Time Streaming Quality through a Multipath Redundant Communication Framework
Koki Ito, Jin Nakazato, Romain Fontugne, Manabu Tsukada, Hiroshi Esaki · 2024
Recently, as networks operate as the infrastructure of modern society, the demands placed on the network by applications have become more complex. In particular, there is an increasing annual demand for high-capacity and low-latency services, including real-time streaming. 5G has been launched to meet this demand, but its stability varies depending on location and time and can only sometimes be considered sufficient. One method to improve communication stability is multipath redundant communication, and much research has been conducted in this area. However, most of this research has focused on TCP-based communication and cannot be applied effectively to real-time UDP streaming. Hence, we propose a multipath redundant communication framework to improve the quality of real-time media streaming communication. Our proposed system is verified using multipath redundant communication and multiple mobile networks from a vehicle moving in an urban area. The experiments use a real-time streaming application based on WebRTC, and the framework significantly reduced packet loss and improved bitrate compared to existing multipath redundant communication systems without interfering with the congestion control mechanisms of the application.