An undisclosed issue with popular content distribution network (CDN) provider, Cloudflare, brought down several popular websites and apps on Tuesday. The service outage began at around noon, affecting services such as brokerage platform Zerodha, communication service Discord, virtual private network (VPN) service NordVPN and more.
About an hour after the outage was reported, Cloudflare announced on its site that the issue has been fixed.
Cloudflare has confirmed that the issue was a technical glitch of the highest critical rating, causing the service network to be disrupted in “broad regions” – suggesting that the outage likely affected apps and sites worldwide. The company further said that the issue impacted “all data plane services” – suggesting that the issue in the company’s servers broke data transfer protocol, which is required for a website to operate.
The outage hit services including that of global cryptocurrency exchange Coinbase, gaming platform Steam, video conferencing service Skype and ed-tech platform Udemy among others.
Outages such as this, albeit not regular, can affect vast swathes of the internet, bringing down various popular websites and apps that are used on a regular basis. In 2021, fellow CDN provider Fastly saw a similar service outage that brought down various popular services such as game streaming platform Twitch, music recognition platform Shazam, online community Reddit, code repository site GitHub, and more.
A CDN operates a vast network of distributed servers, which helps reduce the amount of time that a website takes to transfer data from a server where its information is stored, to its front-end for users to access. Prior to the advent of CDNs, websites would take much longer for data transfers to take place via central servers. A CDN, therefore, drastically reduces the amount of time that a site takes to display information by localizing servers closer to the source region of a user request.
As a result, most popular apps and sites rely on CDNs as one of their network infrastructure providers to operate. Given their general stability, most sites also do not deploy a backup plan due to such infrastructure being quite costly.
For instance, in April this year, a glitch in cloud servers operated by Amazon Web Services (AWS) caused Indian food delivery platforms, Zomato and Swiggy, to face outages at the same time.
At the time, Doug Madory, director of internet analysis at American IT services provider Kentik, had explained that the reason why such outages often take down vast chunks of the internet when an issue occurs is due to the overall cost of such infrastructure, and their general reliability.