Derby and Amazon S3?

So, my JavaOne session on Derby that I’m doing with Dan Debrunner got approved, and I just submitted the slides. In the slides I committed to doing a demo showing how to use Derby on a memory stick. I have a bunch of ideas of what I might do with this, but one that’s got me kind of curious is somehow integrating data stored in the newly announced Amazon S3 (Simple Storage Service)

The problem I’m having is that the S3 schema is so basic – just keys and values. It makes me wonder if I should add a new storage plugin to Derby that runs on top of S3, so that it could be queryable. Or maybe I could find a way to cache S3 data locally in Derby and synchronize it on some kind of regular basis, or on demand. Now, that could be cool. Use S3 as an Internet backup for my database on a stick...

What do you think?