Survey OF Effects Of Cloud Outage Shows How Much Of The Web Runs On Amazon - Forbes

All of the news reports about the Amazon Web Services Outage Friday night mentioned the same few big name companies, Netflix, Pinterest and Instagram. A few tech blogs mentioned Heroku, the well-known Ruby and JavaScript-based cloud application platform. Clearly Amazon and all of the companies effected are working on damage control, which means, for the most part, a tweet or three but no public acknowledgement on their websites.

Just based on what is in the public domainâ€"particularly those tweetsâ€"it is possible to piece together the extent of the impact of the outage, and it was significant. Although most of the feedback I received to my post yesterday about the outage strongly cautioned against using the cloud for anything truly mission critical, this survey will show that many companies are using AWS extensively. Many new companies have business models that only make sense (if, in fact, they do) because of how inexpensive and flexible Amazon’s cloud services are. So before anyone gets too irate about who is responsible for what and how could this have been prevented, let’s remember that most of what is now hosted on Amazon didn’t even exist ten years agoâ€"and we survived!

Beyond the specific outages reported by individual companies, there has been a larger impact on the web as a whole over this weekend. So many “web services” that are called by web pages are hosted on Amazon, for example e-commerce, email, databases, etc. And many of the accounts had to be restarted by the clients themselves after the outage, which fell inconveniently on a Summer weekend. So this is to say that many more sites have been effected than just those who are actually hosted by Amazon’s cloud. Here is a survey of the 24 hour impacts of the outage and how the companies have informed their customers about them.

First, the ones we all know about:

Netflix: No public information available about outage on website. Tweet at 1:33 PM - 30 Jun: ”If you still can’t connect this morning, find your device here:http://bit.ly/LHNOhB, then remove and reinstall the app.  # dcstorm”

Pinterest: No public information available about outage on website. Tweet at 11:05 PM PSD â€" 29 Jun: “We’re back! Our team is continuing to work on some remaining issues that may impact performance. Thanks for your patience and happy pinning!”

Instagram:In the help section of its website, the company posted this message Friday night: “Instagram Is Not Loading. As of Friday evening of June 29, 2012, Instagram is experiencing technical difficulties. An electrical storm in Virginia has affected most of our servers, and our team of engineers is working hard to restore service. Thank you for your support and patience!” Tweet @11:13 AM â€" 30 Jun: “Due to severe electrical storms, our host had a power outage, no data is lost â€" we’ve been working through the night to restore service”

Heroku: Highly detailed public information available about outage on website (see graphic above. Wow!) Tweet @9:55 PM PDT â€" 30 Jun: “Issue: Shared Database Server Offlinehttp://status.heroku.com/incidents/390# …”

In looking around for a ready list of companies whose sites are hosted by AWS, I came across the Case Studies page on their website. Here is my pick of some of the better known names on their long list of (previously?) happy customers:

99Designs: Crowd-sourced graphic and web design marketplace. Tweet at 10:38 AM PDT - 30 Jun: “Site is back up, although with less databases than usual. We’ll see how it goes ^lox.”

Active.com: Online sports community for competitors to register for races and other recreational activities. No reported impact.