How CDNs Work: Edge Delivery, Caching & Performance
A CDN is a distributed network of edge servers that cache and deliver content closer to users. It improves speed (shorter distance), reduces origin load (fewer requests), boosts reliability (distributed capacity), and adds security (edge protections). The real power lies in smart cache policies.
Instead of every customer traveling to the main factory (origin server) for every item, local warehouses (edge PoPs) keep popular items in stock. When a customer orders, they get it from the nearest warehouse. If the item isn't there, the warehouse quickly restocks from the factory and keeps it for the next customer.
1How a CDN Works
Client → DNS/Anycast routes to nearest PoP → Edge checks cache key → Hit: serve immediately. Miss: fetch from origin, cache response, return to client.
2Key Benefits
Faster load times, reduced bandwidth/egress costs, higher availability through distribution, and edge-layer security (DDoS protection, WAF).
3Cache Policy Fundamentals
Cache key design, TTL / Cache-Control headers, and purge strategy determine hit ratio, freshness, and cost.
| Property | Without CDN | With CDN |
|---|---|---|
| Latency | High (full distance to origin) | Low (nearest edge) |
| Security | Basic | Enhanced at edge |
| Origin Load | High | Low (cache hits) |
| Reliability | Single point of failure | Distributed |
Without CDN
Latency
High (full distance to origin)
Security
Basic
Origin Load
High
Reliability
Single point of failure
With CDN
Latency
Low (nearest edge)
Security
Enhanced at edge
Origin Load
Low (cache hits)
Reliability
Distributed
Common questions
- ›“How does a CDN work?”
- ›“What are the main benefits of using a CDN?”
- ›“Explain cache hits vs misses and how TTL affects them.”
- ›“How would you design caching strategy for a high-traffic site?”
What interviewers look for
- Clear request flow (DNS → Edge → Origin)
- Understanding of cache keys, TTL, and purge mechanics
- Balance between speed, freshness, and cost
- Connection to frontend performance (LCP, TTFB)
Short answer (60 sec)
A CDN caches content at edge locations worldwide. Requests are routed to the nearest PoP. Cache hits are served instantly; misses fetch from origin and populate the cache. This reduces latency, origin load, and improves reliability and security.
Detailed answer (senior level)
CDNs use Anycast DNS to route users to the closest edge server. Smart cache policies (keys, TTL, purging) determine efficiency. Benefits include faster delivery, lower bandwidth costs, better availability, and edge security. Trade-offs involve staleness risk and cache invalidation complexity. In frontend, CDNs directly improve LCP and TTFB when configured correctly.
- Thinking a CDN replaces proper origin hosting
- Overly broad or incorrect cache keys (stale/personalized content)
- Using very long TTLs without purge strategy
- Forgetting to set proper Cache-Control headers
- Assuming all content should be cached equally
- ✓CDN = distributed edge cache layer between users and origin
- ✓Cache hits serve from nearby PoP for speed and lower origin load
- ✓Cache key + TTL + purge strategy control effectiveness
- ✓Major benefits: speed, cost savings, reliability, security
- ✓Not a replacement for good origin architecture
- ✓Measure real impact on TTFB, LCP, and origin traffic
- ✓Policy design matters more than just having a CDN