Origin Server
The Source of Truth Behind a CDN
An origin server is the server that holds the authoritative version of a website's content and runs its application. When a CDN is in use, the origin is the source the CDN fetches from whenever it does not already have a valid copy. Fetching this way is known as origin pull. The edge servers keep copies, but the origin remains the place where content is created and updated.
Think of the origin as a main library and the CDN as a set of branch libraries. Each branch keeps popular books, but new titles arrive from the main library. If the main library closes for a short time, the branches can still lend what they hold, as long as their copies have not expired.
What Happens When the Origin Fails
Origin errors are a normal part of running a site. A server may restart, a deployment may break, or a database may stop responding. How a CDN reacts depends on its settings. Many CDNs can be configured to keep serving expired copies while the origin recovers, while others pass the error on to the visitor.
This behavior matters most for static files. A site whose stylesheets and images are still cached at the edge can often appear to work during a brief origin outage, while pages that need fresh data show errors. Setting clear rules for how long expired copies may be used during failures is part of planning a CDN setup.
Why the Origin Address Should Stay Private
Visitors should reach the site through the CDN, not directly through the origin. If an attacker learns the origin's IP address, they can send traffic straight to it and bypass the CDN's protection and caching. A common safeguard is a firewall rule that accepts web traffic only from the CDN's published address ranges.
Keeping the origin hidden is a basic part of using a CDN safely, and it is worth checking after any change of hosting provider. Some networks add a middle layer that reduces repeated origin requests, described in the origin shield article.
Frequently asked questions
Can a website have more than one origin server?
Yes. Larger sites may run several origin servers or regions, with the CDN choosing which one to use or failing over between them. Smaller sites usually have one origin, and that is enough for most purposes. Multiple origins add complexity, so they are usually adopted when uptime or regional demand requires it.
Is the origin server the same as my web host?
Often, yes. If your files and application run on a hosting account, that account is the origin for a CDN placed in front of it. If the CDN is the only place content is stored, as on some static site platforms, then that platform acts as the origin.