Around midday on October 22, a number of top websites all went down, among them including Reddit, Minecraft, Pinterest and Foursquare. The iconic 404 of these sites and many more is a shock, especially considering just how essential these websites are to most users of the web as a source of social media and activity alike.

reddit The cause? The downtime was cause by an outtage in Northern Virginia from Amazon Web Services, one of the world's largest cloud server systems, and host to millions of websites, including some of the internet's most largest websites and services. Amazon Web Services runs thousands of servers around the world, and is considered one of the most robust services to host high-bandwidth content through. Despite this, this is the second major outtage this year, the other last June, when Netflix experienced a major outtage due to servers going down in the AWS system.

While the cause for these outtages can always be up for debate, it's essential to think instead of the implications of this, and a demonstration of just how fragile our robust interenet is. While it's surprising to think of dozens of top 100 websites bring brought down by a single server outtage, it's a fact we have to live with as major companies like Amazon roll out Content Networks like AWS that are a foundation to millions of websites at a time. Just last month, GoDaddy's DNS server experienced a Denial of Service Attack, bringing as many as 48 million websites to their knees. The reality is apparent, as amazing as our web has grown in the last 2 decades, only a handful of servers control the gates that keep our favorite websites alive and well. While AWS has had any otherwise impressive track record, this event is another example to top websites of how essential it is to have contingencies in the event of downtime, even with the world's largest data centers. As of this writing service has been partially restored, and it won't be long until the rest of our favorite websites are back online.

About the Author

I write, I read, and I love Technology

They are only fragile because they do not utilize failover and high availability... I really wonder who their sysadmins are because it seems borderline negligent.

My company website also survived from AWS outages for about 12 hours.

Any suggestion for automatic ec2 backup in another zone?

I think it'd depend on your scale Slinda

We at Total Uptime Technologies have a Cloud Load Balancing and Cloud Failover solution that allows companies to fail from one AWS region to another. The AWS load balancer is great for failing over between availability zones within the same region, but that is as far as it goes.