KEPC-Push: A Knowledge-Enhanced Proactive Content Push Strategy for Edge-Assisted Video Feed Streaming

Explore a 21-minute conference talk from USENIX ATC '24 that introduces KEPC-Push, a novel Knowledge-Enhanced Proactive Content Push strategy designed for edge-assisted video feed streaming platforms like TikTok and Reels. Learn how this innovative approach addresses the challenges of proactive content distribution in the absence of historical access information for recently generated videos. Discover how KEPC-Push utilizes knowledge graphs to determine popularity correlations among similar videos based on factors such as authors, content, and length. Understand the hierarchical algorithm developed to optimize resource allocation across edge nodes with varying capabilities, operating at a regional level to reduce communication distances. Examine the results of trace-driven simulations that demonstrate KEPC-Push's ability to reduce peak-period CDN bandwidth costs by 20% and improve average download speeds by 7% compared to existing solutions.