Critical services, debugging, and global problems: Fastly’s story

Fastly network's status after being back to normal
Fastly network’s status after being back to normal

On Tuesday, a lot of websites worldwide suffered malfunctions that led to some parts missing up to being impossible to connect to them. That was due to a serious problem of Fastly, a CDN (Content Delivery Network o Content Distribution Network) supplier, which means a service that normally allows faster access to a website. The problem started when it was night in the USA with consequences all around the world then it was solved about an hour after being recognized at Fastly. The trouble didn’t last long but show how something going bad in the system of a very important Internet operator can have consequences on a lot of websites. In this case, it was a bug introduced in a recent update that was supposed to be discovered before it caused any damage.

We’re used to thinking about the Internet as a global structure that connects all the world canceling distances. Actually, the physical features of the apparatuses that allow the connection between websites, or other Internet services, and the visitors are crucial. That includes the length of the “road” the signals have to travel. For this reason, for a website with huge global traffic, it can be useful to take advantage of a CDN that offers a number of copies of the website in different parts of the world and connects the visitors to the copy that’s physically the closest to them.

Fastly was founded in 2011 by Artur Bergman, and in the course of the years, it’s become a big name as a supplier of CDN service, which can include security and protection services. Websites from all over the world rely on Fastly and for this reason, on Tuesday, they got struck by a problem that turned out to be global.

It’s normal for the owners of websites with a lot of visitors from around the world to rely on suppliers of CDN services such as Fastly. Tuesday’s incident shows how a fault to an important supplier can cause significant problems at a global level. Of course, it was a peculiar event that seems to have involved all Fastly CDN’s nodes canceling the advantage of their redundancy. However, for websites that belong to public administrations or that offer bank services and the likes, depending on third-party services can turn out to be a double-edged sword.

Fastly is a public company at the American Stock Exchange and its shares had a fall, however, it quickly recovered after the company’s services got restored. A company’s press release explained that a system configuration caused the global fault suffered by its customers. The problem was quickly pinpointed and fixed bringing the situation back to normal.

Fastly didn’t hide behind buzz words but was transparent and honest by publishing an explanation on the company blog. The mistake was acknowledged and an awareness that the company provides critical services was shown. That’s good because it will help improve the company’s software development procedures and consequently the debugging.

Leave a Reply

Your email address will not be published. Required fields are marked *