A Buffer-Based Rate Adaptation Approach for Video Streaming Over HTTP

The demand for high quality multimedia streaming services has increased rapidly in the past decade. Sustaining this growth while maintaining the quality desired by the end users poses several challenges to the service providers. Perhaps the most important challenge is the fact that the underlying networks for these streaming services are fundamentally best-effort, and that all the attempts to offer an acceptable quality of experience (QoE) to the users were necessarily an afterthought. Nevertheless, some service providers have managed to offer very competitive streaming QoE thanks to the well-refined dynamic adaptive streaming over HTTP (DASH) technology. In this work, we propose a buffer-based streaming algorithm for dynamic adaptive video streaming over HTTP. The proposed algorithm aims at jointly reducing quality variations as well as playback interruptions by improving bandwidth utilization. This is done by maintaining the occupancy of the client playback buffer around a level that enables the streaming approach to absorb any possible undesirable network dynamics. Specifically, the client decides the segment bit rate by considering the variations in the video quality, buffered playback time in addition to bandwidth utilization. The simulation results show that the proposed system can maintain the playback process while gradually switching between the available video qualities while enhancing the bandwidth utilization.