CDN / edge cache
Geographically distributed caches that serve static and cacheable content near users.
Problem it solves: Serving large or popular content from a single origin is slow and expensive in egress.
Where it sits: Between clients and the origin load balancer.
When to use
- Read-heavy static assets, media, and cacheable API responses.
When not to use
- Highly personalised or rapidly changing responses with no shared cache key.
Failure modes
- Cache stampede on purge or expiry overloads the origin.
- Misconfigured cache keys leak or fragment content.
Metrics to watch
- Edge hit ratio (%)
- Origin offload (req/s saved)
- Edge latency (ms)
Alternatives
- Origin shield
- Application cache
Configuration in Labs (illustrative defaults)
- Steady-state hit ratio (0–1) = 0.9 — Fraction of requests served from the cache once warm. Misses go downstream.
- Hit latency (ms) = 15 — Time to answer a request that hits the cache.
- Warm-up time (s) = 30 — When the cache starts cold, the hit ratio grows linearly to its steady-state value over this time.
Vendor examples (pattern only): CloudFront, Cloudflare, Akamai, Fastly
In the guide
- Ch. 5 · Caching and Content Delivery p. 1
- Ch. 18 · Design a Video Streaming Platform p. 55
- Figures: 14