views:

43

answers:

2

Is there currently a filesystem agnostic way to have a file in two locations on a network, change one copy, and have just the changed bits (or more likely blocks) synced to the other copy?

It would be awesome to have media files in the cloud, allow another service to read them and me to edit them with their web app, and then only have to save back the delta for time/efficiency.

I'm sure people have done this with proprietary in-house methods (Dropbox claims to only need to sync changes rather than the whole file), but is there a more standardized way to do this? This strikes me as something that would be a VERY useful addition to WebDAV if it hasn't been done already.

I'm pretty much thinking of ZFS/BTRFS style snapshots/backups but over http or some other web protocol that can handle latency and dropped packets.

+1  A: 

Looking for rsync (for example over ssh)?

ChristopheD
Yeah, how did I miss rsync? Using with s3 could be great. Now I need s3 to implement rsync natively so you don't have to fire up an ec2 instance to do it and all the web apps need to start using it so we can have all our files in one place but still use the services we want.
kjs3
A: 

Did you check out Unison? It's rsync on steroids. ;)
Installing on Ubuntu.
How to set it up?
Why Unison and not rsync?

Sudhanshu