an edge server

If a CDN is a chain of warehouses placed near customers, an edge server is one of those warehouses — and the clerk who works there. It is a caching server positioned at the edge of the network, geographically and topologically close to a population of users, that holds copies of a website's content and serves requests on the origin's behalf. CDNs deploy these in clusters at many points of presence (PoPs) around the world, often inside or next to the local ISPs and Internet exchange points where users connect.

When a request reaches an edge server, it follows a simple decision. If the requested item is already in the edge's cache and still fresh, it returns a cache hit — fast, with no contact to the origin. If the item is missing or stale, it is a cache miss: the edge fetches the item from the origin (or sometimes from a nearer parent cache in a tiered hierarchy), stores it locally — a cache fill — and serves it; the next nearby user then gets a hit. Beyond caching, modern edge servers also terminate TLS, compress responses, reuse connections to the origin, and increasingly run small programs (edge computing / edge functions) that customize or assemble responses right there.

Why it matters: the edge server is where the CDN's promise actually happens — it is the nearby copy that turns a 150-millisecond intercontinental round trip into a few-millisecond local one. The distinction to keep straight is origin versus edge: the origin is the authoritative source that owns the real, up-to-date content; the edge holds cached copies and must rely on freshness and validation to avoid serving stale data. The edge is fast and plentiful but borrowed; the origin is the single source of truth and the place changes must ultimately come from.