On Data Wastage When User Becomes Idle in HTTP Adaptive Streaming

HTTP Adaptive Streaming (HAS) has become the fundamental technology of nowadays' online video streaming services to maintain the user's Quality of Experience. Such an ability of the HAS is powered by its Adaptive Bitrate Algorithm (ABR), which focuses on maximizing both the visual and audio quality of the video with regards to the user's network condition. However, not all users equally care about both visual and audio quality concurrently. Our subjective assessment on viewing behaviors shows that, for videos with audio-focused content, most of the participants let the video run in the background and move on to other works for a certain time. As a result, maintaining a high visual quality within this period, which current ABRs are doing, becomes unnecessary and thus causes wastage of data. To tackle this problem, this paper proposes a novel algorithm called Idle-Aware, which adjusts video quality based on the idle behavior of the user. The experimental results showed that our solution not only reduces data wastage but also diminishes the rebuffering events under congested networks.