Yesterday was not a regular day for the Internet. When I tried to check my tweets in the morning, I found that Hootsuite, the browser-based social media management tool, was down. Oh well, I thought, maybe Reddit will get me some interesting links instead. But to my horror, it was suffering from an outage as well.
Then came the straw that broke the camel’s back and convinced me that something, somewhere is terribly wrong: I couldn’t check in to anywhere on Foursquare!
It turned out to be an issue with Amazon EC2, a cheap and scalable cloud service that many, many websites and services depend on for their hosting. On top of the above-mentioned Hootsuite, Reddit, and Foursquare, other sites affected include Q&A site Quora, location-based mobile game SCVNGR, iPad music app Discovr, social media monitoring app Wildfire, and more.
At the time of writing, services to these sites have mostly been restored, although there are evidently some kinks yet to be worked out. For example, over at Reddit you still can’t post if you’re a regular member, you can only read. Amazon said on its status page that it is working hard to restore services to all sites.
One common feature of these sites and services is that they are mostly start-ups. Amazon’s cloud hosting is one of the cheapest available. By renting server and computing space from Amazon instead of buying and building their own infrastructure, these young companies can shift their money to other more pertinent spending.
But as more and more services move to the cloud, an outage at a large cloud hosting provider could bring down a large part of the Web, as demonstrated yesterday by Amazon. This puts a serious cringe in cloud computing’s bid to be a viable competitor to traditional webhosting.
It seems crazy to let an outside party take care of such a crucial part of your site/service: your uptime. But computer downtime is a fact of life, much like taxes: inescapable even if your cloud services are highly reputable.
There’s also a certain danger to letting a few large providers “own” (so to speak) a majority of the Web. The Internet was designed as a resilient network that can route traffic around failed sites, so that the rest stays running. If everyone depends on companies like Amazon to run their sites and services, the Web could well become centralised and a disruption could quickly turn into a major catastrophe.
Has this episode shaken your faith in cloud computing? Let us know in the comments!