CDN Challenges of HTTP-based Low Latency Live Streaming Delivery
summary
Claims that HTTP live can hit glass-to-glass latency under four seconds by combining specific CDN and encoding changes: nginx-based request coalescing and a custom origin that serves incomplete files from memory, cache pre-warming / pre-push of chunks to edges, and HTTP chunk transfer (serve partial segments plus H2 push or QUIC) to avoid waiting for full segments. Also uses per-title bitrate ladders (convex-hull) and a push vs pull transcoding tradeoff; shows measurable reductions in startup and origin request load.
AI-written summary · about
chapters (11)
- 0:00Low latency CDN challenges
- 1:29Glass-to-glass latency
- 2:13Latency in HTTP live streaming
- 3:17Transcoding trade-offs
- 4:45Push vs pull transcoding
- 6:52CDN software optimizations
- 9:03HTTP chunk transfer issues
- 11:35Preloading segments
- 12:40Zero RT protocols and caching
- 14:06Network backbone for low latency
- 15:32Efficiency in data transfer
related talks
19mLow Latency Live from a Different Vantage Point
21mThe challenges of deploying Apple's Low Latency HLS In Real Life
10mEnhancing CDN Performance and Cutting Egress Costs in Large Video Libraries
44mLive Video in a Modern Web Environment
36mCHUNKY MONKEY
17mLive Video Streaming as "tail -f"
33mThe Challenges of Testing Video Playback at Scale
53mCDNs 101: An Introduction to Content Delivery Networks