Tuesday, July 22, 2008

Amazon S3 downtime

You may have heard that Amazon S3 was down for a fairly long time on Sunday (not sure how long, I hear between 3 and 8 hours).

It raises an interesting point: considering what 99% uptime actually means in terms of hours per month, should developers think about redundancy? And what is the best way to go about it? What if you have a partial lock-in by using Google app engine or Jaxer?

My guess is that the open source community will eventually catch up to offering PHP mashups that are similar to popular "cloud" APIs, and that numerous web hosting companies will start adding these mashups as features to remain competitive. It's happened with Fantastico, even Rails is now available in a number of hosts that didn't offer it a year or two ago.

As far as S3 goes, I'm not too familiar with the options. Perhaps one could consider Edgecast? Does anyone know of other providers?

No comments:

Post a Comment