BINARY: A bandwidth variation pattern-aware rate adaptation for HTTP adaptive streaming over 4G cellular network

Recently, HTTP Adaptive Streaming(HAS) has been an essential solution for mobile video streaming to improve the Quality of Experience(QoE). The rate adaption of HAS dynamically switches between different quality levels of the same media content as the available bandwidth changes. In this paper, we propose BINARY, a novel HAS rate adaptation algorithm for the 4G cellular network with a dual character. BINARY differentiates the patterns of bandwidth variation as constant bandwidth fluctuation and instantaneous bandwidth hopping. In the first case, BINARY performs smoothed bandwidth prediction and conservative rate switching to avoid frequent video quality oscillations. In the other case, BINARY performs positive bandwidth prediction and aggressive rate switching to prevent rebuffering and make full use of available bandwidth. We evaluate the performance of BINARY through extensive experiments on a 4G network testbed. The results demonstrate that BINARY achieves improved QoE in higher average bitrate, reduced bitrate switching frequency, and lower rebuffering ratio.