| Subject: | Provide enough flexibility to support alternative common workloads |
From what I can tell, brackup is primarily aimed at symmetrical, chunked
(if possible) backups.
It would be highly useful if there were alternative algorithms available
for alternative workloads.
For example, if I'm backing up my family photos, I don't really care
about strong or any encryption, and I'm happy with archive rules being
in place ("You can add files, but never change them")
The files are all half between a half and four megabytes, so chunking
probably isn't needed. And so on, and so on. SVN fsfs repositories are
made to be rsync friendly, so you can back up each commit as a single file.
Being able to use alternative backup strategies like this could be
hugely valuable. Not only that, but when combined with Target cost
extimates, you might well be able to provide a way to let people see
which backup strategy will result in the lowest cost, or automatically
select one based on the content, and so on.
Anyways, these don't have to be done now (and I highly doubt they should
be) but it would be nice to have room in the API to allow it down the line.
Some common workloads you might want to think about are photo albums,
source code repositories or various types, and various email things.
Anyways, something to leave in the queue a bit longer term.
Or that you explicitly say you will never support. :)