Anticipatory Download Scheduling in Wireless Video Streaming with Uncertain Data Rate Prediction

Video streaming is in high demand by mobile users. In wireless networks the unreliable wireless channel leads to two major problems. Low available data rates degrade video quality and interrupt the playback when a user cannot sufficiently fill its local playout buffer: playback interruptions occur. In contrast, high available data rates cause common greedy buffering schemes to buffer too much data. Such over-buffering wastes wireless resources. Assuming that we can predict future data rates, we plan the quality and download time of video segments ahead. This anticipatory download scheduling avoids playback interruptions by downloading a large number of segments before a drop in available data rate occurs, without wasting wireless capacity by excessive buffering. In previous work we assumed a perfect prediction of future available data rates. Here we assume uncertain predictions, which are subject to errors and are only available for a limited future. We developed a stochastical model for prediction errors and implemented a new scheduling algorithm that is aware of prediction errors. Our results show that anticipatory download scheduling is feasible even with uncertain predictions and reduces playback interruptions while delivering a high video quality.