Answers are grounded in this paper and cite their sources.
Demonstrates enhanced video seeking in streaming, suggesting advancements in user experience through prefetching techniques.
Streaming video dominates traffic on today's Internet. For many content types, users do not watch the entire video, but rather skip around within the video. When a user skips to a time not yet buffered by the client software, current systems stall the video playback while fetching the relevant chunk. Meanwhile, end-user bandwidth continues to increase with increased deployment of fiber-to-the-home and 5G services, as CDN appliances inside of ISPs drive down round-trip time (RTT). This paper introduces StallFreeSeek (SFS), which can provide nearly stall free seek for users by proactively prefetching the latency-sensitive portion of future video chunks without increasing buffer waste. SFS carefully rechunks the video such that the bandwidth overhead for prefetching is small and the buffer can be refilled after user seek without stalling video playback. We evaluated SFS using various video genres, different models of user seeks using over fifty thousand simulated video sessions and nearly two thousand real-world video sessions. In these evaluations, SFS consistently outperforms Dash.js in QoE, stall time, and buffer waste. For example, in our ABR experiments on real-world network performance traces, SFS improved QoE by 25 percentage points, reduced seek-related stall times by 79%, while reducing buffer waste by 44%.
No takes yet. Share an insight, caveat, or question.
Zhu et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: