Towards User-Oriented Live Video Streaming

With the development of innovative network infrastructure and increasing bandwidth availability, live video streaming is emerging as an attractive application for end users and the industry. Due to the heterogeneity in user resources as well as bandwidth fluctuations, we argue that, to provide high quality-of-service, it is desirable to tailor (on-the-fly) the bit rate of video streams, according to user download bandwidth availability. We believe that at least two benefits can be obtained by such an approach: (1) delay avoidance when a user´s bandwidth drops below the (current) video playback rate; (2) provision of better quality of video streams, when a user´s bandwidth increases, e.g., if a video stream with a matching bit rate is unavailable at the server. To this end, we propose a user-oriented video streaming system, uLive, that can on-the-fly adjust the streaming bit rate in order to match available user bandwidth. Our experimental results indicate that uLive is adaptive to user heterogeneity and robust to bandwidth fluctuations, with the goal of maximizing quality of the video received by the user.