rdiff-backup can make backups while conserving bandwidth, which is probably a great idea on the whole. It also makes incremental backups, and the home page promises file recovery over previous backups too.
I didn’t delve that far into it, but I do have a little to show for my effort:
My hope there was to show that rdiff-backup’s product is not only a mirror image of the source, but also includes data on what changed between runs. It might be a little difficult to follow; trust me if it’s not obvious.
Compared to a straight rsync, I can see where this would be preferable, if it conserves bandwidth and can offer access to past backups as well. I usually just refresh my archives with a simple
rsync -ah --progress --delete, and there have been times I wished I could step backward once or twice in history.
On the other hand, this is very clean and straightforward, without a lot of the wrangling that I’ve seen in some other console-based backup tools. Given the need — such as a large-scale networked system — I’d definitely think this over as an option. 😉