Edge Network
Definition
An edge network is the distributed layer where requests are actually answered, and it is what a CDN is physically made of. Rather than one data centre serving everyone, the provider operates dozens or hundreds of locations and routes each visitor to a nearby one; the word edge means the outer boundary of the network, the last stop before the visitor's own connection. Each location is a point of presence, and the whole set of them is the edge network. It does more than store files. Compression, TLS termination, HTTP/2 support, request filtering and increasingly small pieces of code in the form of an edge function all run there, which is why a modern platform describes its features as happening at the edge rather than on a server.
Why It Matters
Where the work happens decides what a visitor waits for. Terminating the encrypted connection at a node 40 kilometres away rather than a data centre 8,000 kilometres away turns a handshake costing 300 milliseconds into one costing 20, and that is before any content moves. Once files are cached locally, time to first byte for a repeat visitor typically lands in the 20 to 50 millisecond range instead of several hundred, which is the difference between a page that feels instant and one that feels like a website. There is a resilience argument as well: when one location goes offline, traffic shifts to the next nearest, so what would once have been an outage becomes a few tens of milliseconds of added latency.
How It Works
Routing is usually done with anycast: every node announces the same IP address, and internet routing naturally delivers each request to whichever announcement is closest in network terms, which is often but not always closest geographically. DNS-based steering achieves something similar by returning different addresses to lookups from different regions. The node that receives the request terminates TLS, checks its cache for the path, and either serves a stored copy or fetches from the origin and stores what it gets according to the Cache-Control header on the response. Nodes do not share a single cache — each keeps its own — so the same file is fetched separately for each region the first time it is wanted there. Providers soften that with a shield node that all the others pull through.
Real-World Example
A conference publishes its schedule as a small static site on 99helpers at torran-conf.99helpers.site, and attendees open it on phones in the venue, in hotels and on the way home. Each of them is answered by whichever location is nearest, so the page appears in well under a second whether they are in Manchester or Melbourne. The organiser uploaded a folder and did nothing else; on 99helpers the distribution is simply how files are served. The edge analytics counting those visits are collected in the same places, without cookies.
Common Mistakes
- ✕Counting advertised locations as a quality measure — a provider with 40 well-connected sites often beats one claiming 300 thinly provisioned ones
- ✕Expecting one node's cached copy to help another — caches are per location, so the first visitor in each region still pays for a fetch from the origin
- ✕Assuming everything runs there — an uncacheable or personalised response is passed straight through to the origin, and the nearby machine only adds a hop
- ✕Testing from one country and calling the result global — the only meaningful measurement is from several regions, ideally the ones your visitors are actually in
Related Terms
CDN
A content delivery network: a set of servers spread across the world that keep copies of your files and answer each visitor from somewhere close to them. It is the service; the machines it runs on are its edge network.
Point of Presence
A single physical location in a delivery network — the racks of servers in one city that answer requests from that region. Shortened to PoP, it is the smallest unit an edge network is counted in.
Cache Hit Ratio
The share of requests answered from a cache rather than passed back to the origin, written as a percentage. It is the one number that tells you whether your caching setup is doing anything.
Latency
Latency is the delay between sending something and it arriving — usually quoted as round-trip time, the milliseconds for a packet to reach a server and come back. It is a separate thing from bandwidth, which is how much data fits through per second once the first packet lands.
Edge Function
A small piece of code that runs on a delivery network's own nodes, near the visitor, rather than at one central server. It suits redirects, headers and routing decisions — not heavy work.
Put a file online in seconds
Drop in a document, an image, a page or a whole static website and share the link — free, with no build step and no server to set up.
Host a file free →