Server-driven rate control for adaptive video streaming using virtual client buffers

Abstract:

Adaptive video delivery approaches perform video rate selection based on the streaming client's throughput estimate. In practice, the accuracy of the estimated throughput is limited due to feedback delay and the unawareness of the dynamics of the underlying HTTP/TCP transport layer. As a result, streaming applications employ large playback delays in the order of tens of seconds so as to maintain continuous video rendering with good quality and bandwidth utilization. In this paper, we introduce DASP, an advanced video rate adaptation, to achieve a low-latency adaptive video streaming. The key components of our solution are a server-side mirroring of the streaming client's buffer, which provides a low-delay feedback for the video rate selection, and a hybrid rate adaptation logic based on goodput and buffer information, which stabilizes the adaptive response to the dynamics of transport layer. We demonstrate the performance of our rate selection algorithm by evaluating the stability of the receiver buffer under low-latency and high-definition adaptive video streaming with variable bit rate encoding over an emulated wide-area network link. The results also show that our approach is promising and applicable for dynamic live streaming with a playback delay as low as the chunk duration.