Netcraft has a great story about Amazon downtime. For what looks like roughly a day, their Alexa portal experienced extremely high latencies. Perhaps we should give them a new logo to represent their downsides:

I’ve always been against third-party web services solutions. What happens when they go down? It’s not a matter of if they go down, but rather when. No matter how many different locations you have, how many servers in your cluster, you will eventually experience a natural disaster, and lose connectivity. Until there’s some better way to maintain uptime, not prone to natural disaster, failure, or human error, you’re better off running your own systems.

4 Responses to “Amazon S3 Goes Down”

1. cjcm says:

   11/22/2006 at 2:24 am

   So if you were in Amazon’s shoe, how would you architect the solution for Amazon?

   BTW, nice new logo for Amazon, I would have replaced the arrow with another, one pointing right down

   Reply
2. bob says:

   4/15/2007 at 4:19 am

   “I’ve always been against third-party web services solutions.” Have you ever worked in enterprise IT? Jeez.

   Reply
3. Elliott Back says:

   4/15/2007 at 3:21 pm

   Yes, I work in Enterprise IT.

   Reply
4. Chris says:

   1/21/2008 at 9:59 am

   Absolutely wrong. When done properly, third-party web services and hosting solutions can be extremely reliable and cost-effective. Many of our hosting and web clients are on a great hosting platform that have redundant backup and daily or weekly backups to S3 storage. We weren’t down a single time last year.

   By the way — regarding this web form — what is the point of a custom CSS style to make the text I’m typing lighter? The goal is to see what you type.

   This is a good example where design and functionality go separate ways.

   Reply