Block 3. Delivery and CDN

How a CDN delivers video from origin to edge — why segments cache so well, and how the offload ratio decides your streaming platform's monthly bill.

Why one CDN is a single point of failure and a pricing trap, and how multi-CDN architecture and content steering route each viewer to the best network.

How CDN cache keys decide a cache hit, why a per-viewer token wrecks your hit ratio, and how tokenized URLs protect video without breaking the cache.

Why a live premiere melts an unshielded origin, how origin shield consolidates requests to cut origin load, and how to design failover origins for OTT.

How CDN egress cost works for streaming: per-GB tiers, the 95th-percentile trap, commit deals, and the four levers that cut your monthly CDN bill.

How distance, CDN regions, anycast routing, and data-residency law decide whether a global streaming service is fast — and legal — in every market.

Low-latency HLS and LL-DASH cut live streaming delay from 30 seconds to 2–5. See where the latency budget hides and what each second really costs.

A live premiere creates a thundering herd that can melt your origin. Size the spike and survive it: origin shield, multi-CDN failover, degradation.

See which viewers are buffering and where — region, CDN, ISP — before they complain. Delivery observability with CMCD, CMSD, and per-slice dashboards.