3 Comments

  • Don't fear the cloud, just build some redundancy into your systems! Plan for the cloud (or ANY resource, for that matter) to be unavailable for a while.

  • Yeah, I'm not sure all the gloom and doom is warranted. It's like any other component in a distributed system, and I think the appropriate move is to plan for redundancy. If the server I'm renting dies, this isn't that different.

  • Except that with S3 what are you going to do, build your own S3 and install it on your servers? How do you make your system reduntant without building everything yourself that you were trying to avoid building by using the services?

    IMO, Amazon needs some local APIs or local instances you can install in the case of a failure... otherwise, you don't really gain much in terms of dev time, because you'll have to build a hell of a lot of infrastructure to handle the cases where their services are down.

Comments have been disabled for this content.