On February 28th, 2017, a “cockup at Amazon facility in Virginia, US” caused its S3 cloud storage to fail.
This fault led to “major websites and services - including Imgur, Medium, Trello, Yahoo webmail, the Docker Registry Hub.. falling offline, losing images, or left running like treacle. Basically, that outage “knackered half the internet, it seems, because it all relies on S3 to store data online”.
Just to stress: this is one S3 region that has become inaccessible, yet web apps are tripping up and vanishing as their backend evaporates away.
AWS, for some reason, insists this isn’t an “outage” but rather a case of “increased error rates” for its most popular cloud service. Infuriatingly, the status dashboard for AWS shows all green ticks at time of writing, despite what feels like a chunk of the internet going missing as a result of the downtime.
Less than five months ago, Yahoo reminded the whole world one major reason, that is privacy and security, why everybody needs really personal alternatives to centralized web service. This Amazon outage points out the other: resilience. It’s ridiculous that a fault in ONE service provider, however big, knocks half the internet down. If everybody had a really usable alternative to Yahoo, Gmail, Facebook and similar services, that could run on any computer she wanted, from reliable community hosting providers to custom computers at home, no single fault or provider may cause “disservices” at that scale. Amazon just joined Yahoo in the list of real world examples of why we all need a percloud now.