I guess it depends on your alternatives. Certainly commodity hard drives are far cheaper, but enterprise level SAN/NAS storage is much more.

I’m going to keep an eye on this. We recently invested in 20 blades for our primary and secondary sites, and a 50k SAN, so I don’t think we’d use it this round, but in the future I can see where this would be very handy.

How long have you been using it?

I use S3 for log archival, and soon I’ll be moving database backups (old ones) to S3 as well. Other things might migrate there over time as I become better acquainted with the service. I’m still not confident enough to host anything “live” there, but I know people are doing that.

I started out by diving head first into Hadoop, S3, EC2… the works. What I found is that it requires you to really immerse yourself in the ways of AWS. It’s a great service, but there’s a lot of stuff that isn’t done – mostly in the area of administrative tools. I also had a couple of conceptual problems relating to failover/redundancy/architecting-for-failure within the EC2 service environment, and I had so many other things on my plate (still do, unfortunately) that this has been put on the back burner for the time being.

Things are progressing rapidly in EC2-land. Soon we’ll have persistent storage, which actually solves a number of the other issues I had with EC2 (via hacks that would rely on persistent storage), and IP addresses that are at least somewhat predictable. People are also writing better tools to manage all of this stuff, and there’s more reading material about how to make running services in that environment a less sanity-depleting experience.

What sort of things do you use AWS for? Are you using S3 and EC2?