A Server-side Dynamic Bitrate Generation for Live Streaming.

Zhang Dongyan,Chongwei Sun,Weihua Li,Yuanjun Xu

DOI: https://doi.org/10.1145/3603781.3603815

2023-01-01

Abstract:In live streaming, the adaptive bitrate algorithm is an effective way to improve user's quality of experience (QoE). However, the majority of existing adaptive algorithms still have problems of performance degradation in a network environment specifically. When players download chunks of video, it is difficult for the server that provides fixed bitrates to provide suitable chunks for most users. In order to allow the bitrates generated by the server to adapt to the changing network environment, we propose a server-side dynamic bitrate generation algorithm that determines the bitrates to be generated according to clients' bandwidth information. In order to reduce the complexity of the algorithm when a large number of users are online, a clustering algorithm is also used to classify the users’ network environments. Additionally, the client selects the bitrate by considering the bandwidth utilization, difference in quality, and buffered playback time. In order to test our dynamic bitrate generation algorithm, we designed an emulator that supports simulation of multi-user online activity, and our simulation results show that compared to a fixed bitrate schema, our dynamic bitrate generation algorithm can improve user's QoE in several different network environments.