I thought I’d finally give this a go to complement my USB HDD backup strategy (ok at least one of them is kept off site, but it isn’t updated on a regular schedule).

It was dead easy to set up an S3 account via the AWS (Amazon Web Services) console.  I tried DragonDisk on Windows, and whilst it was OK for small file volumes, it lacked speed when synchronising larger (40GB, ~30,000 files).  It is still a useful tool for small file volumes, and provides a decent GUI.  I hadn’t even started on my email backups yet and it immediately struck me this wasn’t going to be viable, but seeing as everything is stored on my Sheevaplug, it made more sense to backup directly from this to Amazon’s S3.

I found s3sync which runs on Ruby – this was very easy to configure using this guide.  If I turn off the MD5 checking (it works on last modified date/time instead), then syncing is much much faster – more akin to the traditional rsync procedure.