Phase 1 — Clarify & Scope
Interviewer: Design a Content Delivery Network — think Cloudflare or AWS CloudFront. What do you want to clarify?
Candidate: A few clarifying questions:
- What types of content — static assets (images, JS, CSS), video streaming, or dynamic API responses too?
- What's the geographic footprint — global, or specific regions?
- What's the traffic volume — requests per second and total bandwidth?
- Cache invalidation requirements — how quickly must content updates propagate?
- Do we need edge compute (running logic at the edge) or purely caching?
Interviewer: Primarily static assets and video streaming. Global footprint — 450+ Points of Presence (PoPs). Peak traffic: 10 million requests/sec, 2 Tbps bandwidth. Cache invalidation must propagate globally within 60 seconds. Edge compute is a stretch goal.
Candidate: Let me size the PoP infrastructure:
With typical peak-to-average ratio of 3×: each PoP must handle ~66K req/sec and ~14 Gbps at peak. A tier-1 PoP (major city) might handle 10× that — 220K req/sec.
Target cache hit rate: 90%+ (only 10% of requests reach origin):
Interviewer: Good. What are the core requirements?
Candidate:
- Route user requests to the nearest PoP (<…
In the full tutorial
- Phase 2 — High-Level Architecture
- Phase 3 — Deep Dive & Evolution
- Phase 4 — Robustness & Operations