How CDNs Work: Caching, POPs, and Faster Websites

When you load a website, you expect it to appear instantly. But what happens when that website's server is thousands of miles away? This is where Content Delivery Networks (CDNs) become essential. A CDN is a geographically distributed network of servers that cache and deliver content from locations closer to users, dramatically reducing load times.

The Geographic Problem

Imagine you're in New York trying to load a website hosted in Singapore. Your request travels roughly 10,000 miles each way, taking hundreds of milliseconds just for the network round trip. This latency creates noticeable delays, especially for image-heavy sites or applications requiring multiple server requests.

A CDN solves this by placing copies of content at strategic locations worldwide called Points of Presence (POPs). Instead of traveling to Singapore, your request goes to a nearby POP in New York, reducing latency from 300ms to perhaps 10ms.

How Caching Works in Practice

Let's walk through what happens when you request an image from a CDN-enabled website:

1. First Request (Cache Miss): You request example.com/logo.png. The CDN's New York POP doesn't have this file yet, so it fetches it from the origin server in Singapore. This takes the full round-trip time, but the POP saves a copy locally.
2. Subsequent Requests (Cache Hit): When the next user in New York requests the same image, the POP serves it directly from local storage. No trip to Singapore needed.
3. Geographic Distribution: The same process happens at POPs worldwide. Users in London, Tokyo, and São Paulo all get fast responses from their nearest locations.

Cache Headers Control Everything

CDNs use HTTP headers to determine caching behavior. The most important is Cache-Control:

Cache-Control: max-age=3600

This tells the CDN to cache the content for 3600 seconds (1 hour). When this Time To Live (TTL) expires, the CDN checks with the origin server for updates before serving the content again.

Other important headers include:

• ETag: A unique identifier that helps determine if content has changed
• Last-Modified: Timestamp of when the content was last updated
• Cache-Control: no-cache: Forces the CDN to validate with the origin before serving

Static vs Dynamic Content Challenges

CDNs excel with static content like images, CSS files, and JavaScript. These files rarely change, making them perfect for long-term caching with high TTL values.

Dynamic content presents challenges. Consider a social media feed that updates constantly. You can't cache it for hours because users would see stale posts. Solutions include:

• Short TTL values: Cache for minutes instead of hours
• Edge computing: Run application logic at POP locations
• Selective caching: Cache page templates while fetching fresh data via API calls

Real-World Performance Impact

The performance improvements are measurable. A typical website might see:

• 50-80% reduction in page load times for international users
• Reduced server load on the origin, as 80-90% of requests hit the CDN cache
• Better user experience during traffic spikes, as CDNs distribute the load

Popular CDN providers like Cloudflare, Amazon CloudFront, and Fastly operate thousands of POPs globally, ensuring users worldwide get fast content delivery regardless of where the origin server lives.

What's Next

Understanding CDN basics opens the door to more advanced topics like edge computing, where application logic runs at CDN locations, and modern caching strategies that balance performance with content freshness. We'll explore these concepts and dive into specific CDN configuration examples in upcoming posts.