Cloudflare outage highlights centralised internet problem

When the internet started, one of the key principles was that of adaptive routing, where packets of information (emails, web pages, etc) could be retrieved and delivered via an almost unlimited number of routes. This meant that when an outage occurred or a computer stopped responding, routing tables would automatically be updated and data flows would continue (this is an oversimplification, so ask if you would like to know more).

However, there are now a number of very large service providers on which the internet relies heavily, and if one of these fail, then it can have wide-ranging results.

Cloudflare is a US web performance and security company, which markets itself as a global network, and they had an outage last week. It is reported that, as a result Minecraft, Steam, Amazon Web Services, Shopify, Skype, UPS all had problems.