The Day the Internet Stood Still
Something happened to Amazon yesterday. A thing that the cloud community never would have thought was possible. So unlikely that even the elite hacker group Anonymous (who if you will remember, wanted to take down every website that hindered Wikileaks) shied away from attempting. No one would have ever guessed.
Well, not the whole thing. But one of its server centers did. And on that server happened to be the hosting spaces for over 100 different websites. Of all of those websites, somewhere is the ballpark of 145 sites were affected by the outage. And they weren’t the little sites either. Hootsuite, SCVNGR, Quora, Reddit, AintItCoolNews and Foursquare were just a few of the affected.
This is big news. In fact, it’s even bigger than some are painting it to be. Larger sites, like Facebook, Twitter and even Google, have gone down before. But a large-scale multi-site takedown is just something we don’t see. The last incident that comes anywhere close to comparing to this was when Wordpress went down for a total of 110 minutes in February 2010, taking most registered Wordpress blogs down with it. But that’s 110 minutes, not almost a full day of downtime most Amazon EC2 websites experienced.
Simple takeaway: The cloud is more vulnerable than you think. And with Amazon and a number of other cloud computing providers continuing to expand their business, more websites are being hosted in the same location. Meaning something like this is bound to happen somewhere again. Maybe not anytime soon for Amazon. I’m sure they’re going to implement the proper redundancy systems and programming to prevent future large outages. But for others, this is just an eventuality.
And it’s only going to get worse as time goes by, when something like this happens again. More and more businesses are entirely web based, which means downtime equals revenue lost. For example, in Q4 of 2010 Amazon reported revenues of around $14 million per day. Just one day of downtime would have hit both the company’s sales and its stock prices hard.
The world is changing. Crises are becoming more digitalized. And when a major website experiences some downtime, it makes the news. Consumers get angry. Communications professionals handle this as a disaster response. And for those that are publically traded, stocks go down. It’s basically an all-encompassing package of bad news.
But it can be minimized. In fact, yesterday it was minimized for most sites, because they did 3 things right:
- They had a customized 404 page visitors were directed towards
- The 404 page was open and transparent, explaining what the problem was
- Some put a tongue in cheek page in place, containing either light-hearted or humorous language, sometimes accompanied by a cutesy visual
This could have been a messy crisis. Finger pointing, heated yelling, lack of communication. But it actually wasn’t. The web community as a whole seemed to take on more of a joking or understanding tone (some blaming the incident on fictitious Skynet). Some were angry, but most weren’t. Maybe it’s because we’re used to seeing a 404 error page every once in a while. Or maybe it’s because we understand that technology sometimes just acts up.
Let’s just hope that next time, and there will be a next time, things don’t turn out uglier.