You may have noticed that earlier today, a large portion of the Internet broke down. Like an episode of South Park, major sites including Reddit, Twitch, Amazon, Spotify, BBC, CNN, and Tom's Guide suddenly stopped functioning altogether or had problems with pages not displaying properly.

So what exactly happened? It was content delivery network (CDN) provider Fastly. Here's what you need to know.

A CDN is a group of servers distributed in different geographical locations. They work in concert to ensure fast delivery of Internet content and have become quite important services; Fastly is one of the largest of them, serving a huge number of major websites.

CDNs are a kind of middleman, and Fastly itself operates an "edge cloud" designed to do very much to improve and optimize the end user browsing experience. This includes placing servers as close to the user as physically possible to reduce load times, allow services to handle unexpected traffic spikes, and protect against denial-of-service attacks.

Obviously, this means that if Fastly goes down, users will not be able to access the web content created by its clients.

Fastly's status page first noted the problem at 5.58 EST and promised to investigate further. The problem was identified at 6:44 EST, a fix was implemented 13 minutes later, and the disruption appears to be over. Around this time, affected websites and services slowly began to come back online.

Unfortunately, the problem was not localized, and Fastly reported that this was a global outage, affecting locations from New York to Sydney and everywhere in between.

Affected services include us at Tom's Guide, Reddit, Amazon, Twitch, Etsy, PayPal, Venmo, Giphy, Target, eBay, BBC, CNN, Squarespace (and Squarespace hosted websites), HBO Max, Vimeo, Spotify, Hulu, and other streaming services.

The Fastri has now confirmed that the outage was caused by a flaw in the service configuration, which has now been disabled.

We usually like to think of the Internet as an intangible platform floating in the cloud, but it is completely dependent on physical hardware to operate properly. If a problem occurs with a piece of this hardware or the software that runs it, a failure can and has occurred. It does not matter if it is a server farm, a local Internet exchange point, or something else entirely.

Something similar happened last year when Cloudflare, a major DNS service, went down for about 25 minutes. The problem turned out to be caused by a faulty router somewhere in Atlanta, misrouting traffic to places it shouldn't have gone.

Thankfully, the problem has been identified and fixed, and your favorite websites and services should be back soon.