What Is a Content Delivery Network?

Content Delivery Network Explainer

A content delivery network, or CDN, is a distributed set of servers that delivers web content from locations closer to users than a site’s central origin. A browser still requests the site’s normal hostname, but DNS, anycast routing, or another traffic-steering method directs the connection to an appropriate edge location. If that edge already has a valid copy of the requested object, it can answer immediately. Shorter physical and network distance often reduces latency and keeps repeated traffic away from the origin.

The first request for an uncached object is a miss. The edge forwards a request toward the origin or an intermediate shield, receives the response, and may store it according to cache rules. Later compatible requests become hits until the object expires or is evicted. HTTP headers define freshness, revalidation, and whether shared caches may store a response. Cache keys may include the path, query, selected headers, device variant, or other factors. A poorly designed key can mix content that should be separate or fragment the cache into many rarely reused entries.

CDNs are especially effective for images, scripts, stylesheets, downloads, and video segments, but they can accelerate dynamic traffic too. An edge may reuse connections, terminate encryption, compress responses, optimize protocols, and carry requests over a tuned private backbone to the origin. Some platforms execute application logic at the edge or cache carefully selected API responses. Personalized or sensitive content requires explicit rules so one user’s response is never served to another.

Content changes create a consistency problem. Versioned filenames allow a new object to coexist with an old one while cached copies expire naturally. Purge or invalidation requests can remove an object sooner, but propagation takes time and may be rate-limited. Revalidation asks the origin whether a stale copy has changed and avoids retransmitting the body when it has not. Operators choose freshness periods based on how quickly updates must appear and how much origin traffic they can accept.

The edge also absorbs attacks and failures. Large networks spread traffic across many sites, filter malicious requests, enforce rate limits, and hide the origin address. Web application firewalls and bot controls can run before traffic reaches application servers. Origin shielding reduces the number of independent cache misses that converge on the backend. These features improve resilience but do not replace secure application code, correct access control, backups, or capacity planning.

Choosing a nearby edge is more complex than geographic distance. Internet routing policy, provider peering, congestion, edge load, and service availability affect the fastest path. A user may be sent to a slightly farther site with a better network connection. Health checks and real-time mapping systems shift traffic when an edge or origin is impaired. TLS certificates and configuration must be deployed consistently across the fleet because the edge represents the site to the browser.

A CDN is therefore a controlled layer between users and origin infrastructure. It steers connections, serves reusable objects from edge caches, forwards misses, applies protocol and security functions, and refreshes content under explicit rules. Its performance gain comes from proximity and reuse, while its operational gain comes from distributing load and failure handling. The origin remains authoritative, but most users may receive the majority of bytes from an edge they never know by name.