We know now that the major Internet outage that happened on Tuesday June 8th was caused by an issue with Fastly, a company that offers cloud computing services to various websites across the Internet.
One of those cloud services Fastly offers is a CDN, which stands for Content Delivery Network. And that is what went down this week, taking large parts of the Internet with it.
A CDN allows websites to store copies of their webpages (and other resources such as images and videos) across various web servers located around the world. These copies are called a cache. And when someone visits a website that had a CDN, is simply directs them to the cache on a web server located close (geographically) to the visitor, so as to make the webpage load faster, as well as prevent the website’s own servers from becoming overwhelmed if there is high traffic.
Lots of popular websites employ Fastly’s services, including Reddit, GOV.UK, the BBC and many media outlets including CNN and the Financial Times.
Sponsored Content. Continued below...
But when the CDN goes down, those websites go down too. And Fastly’s CDN went down this week.
But how did it happen?
According to Fastly, it was down to an undetected software bug that was released a handful of weeks earlier that would be triggered when a Fastly customer changing their configuration to a specific setting under a specific circumstance. So in effect, the software bug was essentially a ticking timebomb waiting for those set of circumstances to occur. Which they did, on Tuesday.
That’s as much detail as Fastly offered, and they didn’t explain what those configurations or circumstances were, or if any customer could have potentially triggered the software bug.
Sponsored Content. Continued below...
The time from the outage first occurring to a fix being implemented was just under an hour, Fastly claims.
The outage has demonstrated once again the perils of relying too heavily on single companies to keep the Internet working. In 2020 Amazon Web Services (AWS), another company that offers cloud computing services, suffered an outage that resulted in numerous websites failing to load or function correctly.
Fastly have said in a summary post that they will investigate fully how the software bug was undiscovered and will work to ensure that it does not happen again.