What are content delivery networks and why big websites crash together

You don’t hear much about content-delivery networks, or CDNs, until they stop delivering. A global outage of major websites on June 8 that lasted about an hour was caused by problems at the San Francisco-based Fastly Inc. It took down websites including the New York Times, Bloomberg News and even the UK government.

What is a content-delivery network?

It’s a kind of bridge between a website or app and a user, helping to push data quickly around the internet on behalf of some of the most popular online companies. CDN providers do that by hosting multiple servers and directing people who call up a web page to the nearest one geographically, rather than sending them back to the origin. Say a company’s website is based in the US and gets traffic from France, it’s possible a CDN provider has a server in France that will direct users there. 

Who uses them?

Such networks are popular with large, high-traffic websites and those offering big files to download. The list of those affected by the June 8 outage included Amaz­on, Spotify, Twitch, Shopify and Etsy, according to Down­detector, a website that tracks service outages across the internet. Other major CDN providers include Cloudfare, Akamai and MaxCDN.

Why use them?

Along with speed and the greater ease of serving more customers simultaneously, CDNs are better placed to deliver content such as high-resolution video without disruption. They can also divert traffic to different servers if demand is high or there’s a sudden spike, allowing websites to keep going when they’re under strain. On the other hand, they are costly, do not have servers everywhere and, as evidenced by this latest episode, mean that companies are putting the fate of their websites in the hands of an outside party.

What happened with Fastly?

CDNs don’t fail very often but when they do, it can be spectacular. There are many similarities between this outage and an issue with rival Cloudflare last year. Cloudflare’s problems arose because — in simple terms — the company’s engineers tried to re-route internet traffic and everything exploded. Since website traffic is routed through a CDN’s servers, when the servers break so does everything else. These issues are also hard to prevent, and often happen when companies need to update their systems. Fastly attributed the incident to “a service configuration that triggered disruptions across our POPs (points of presence) globally”.

Was this a hack?

There is no evidence to suggest Fastly’s issues were the result of a malicious cyberattack. By contrast, all website system administrators know that network outages and downtime can happen, no matter the size of their hosting platform.