What is a CDN?
A CDN is a distributed network of edge servers that cache content closer to users. We cover edge caching, anycast and how CDNs peer with ISPs.
Last updated:
A CDN (Content Delivery Network) is a geographically distributed network of edge servers that cache copies of content and serve it to users from a location as close to them as possible. Instead of fetching every file from a single origin server, the client receives content from the nearest edge node (PoP). This cuts latency, reduces load on the origin and improves resilience to traffic peaks and attacks. A CDN rests on edge caching, traffic steering (usually via anycast or DNS) and dense peering with internet service provider networks.
Edge caching and the origin server
On the first request an edge server fetches an object from the origin and stores it in its cache, so later requests are served from the cache (a cache hit). The lifetime of the copy is set by HTTP headers, chiefly Cache-Control and Expires, while ETag and Last-Modified validate freshness via conditional requests. Well-tuned cache rules let the edge serve most requests and offload the origin.
- Static content (images, CSS, JS, video) caches for long periods and reaches a high cache hit ratio.
- Dynamic content is served from the origin or accelerated through kept-alive connections and path optimization.
- Cache-Control, ETag and Last-Modified headers govern the lifetime and revalidation of copies.
- TLS is often terminated at the edge, which speeds up the handshake closer to the user.
- Cache invalidation (purge) forces a fresh fetch of an object.
Anycast and steering to the nearest node
To steer a user to the nearest node, CDNs use two main methods. In the anycast model the same IP address is announced over BGP from many locations, and internet routing delivers packets to the topologically nearest PoP. In the DNS-based model the user's resolver receives the address of a node chosen by its location. Both approaches give lower latency and automatic failover when a node goes down.
How a CDN peers with internet providers
A CDN's effectiveness depends on how close, in network-path terms, an edge node sits to the user's network. That is why CDN operators peer heavily with internet providers at public internet exchange points (IXPs) and over direct PNI links, and some place caches inside the ISP's network (embedded caches, such as the Google GGC or Netflix Open Connect models). Peering shortens the path, removes transit cost for local traffic and yields predictable, low latency.
CDNs and AS202520 SkyPass
CDNs depend on dense peering, and AS202520 SkyPass supplies that connectivity layer: BGP peering and remote access to the Polish internet exchanges THINX, TPIX, WRIX and 1-IX, with PoPs in Warsaw and Wrocław. An ISP or hosting provider can use us to shorten the path to CDN nodes and embedded caches, and where peering does not cover a given resource, fill the gap with IP transit.
Frequently asked questions
How does a CDN differ from an origin server?
The origin is the source server holding the original content. A CDN is a layer of distributed edge nodes that cache copies from the origin and serve them to users from the nearest location, offloading the origin.
Does a CDN require anycast?
Not always. Some CDNs steer traffic with anycast (the same IP address announced from many locations in BGP), others with geolocation-based DNS. Both approaches route the user to the nearest node.
How does a CDN shorten the path to my ISP network?
Through peering at IXPs, direct PNI links and caches embedded inside the operator's network. The closer a node sits topologically, the lower the latency and the less transit traffic.
What controls how long content is cached?
HTTP headers: chiefly Cache-Control and Expires set the lifetime of a copy, while ETag and Last-Modified let the edge validate freshness with a conditional request before re-fetching.