It's impossible to do a real rsync command with cloud storage b/c you can't replace pieces of files. However, you can check for whether a file is likely to be the same and not transfer it.
There's also the concept from zsync which performs all rsync computation on the client side and stores it in a metadata file on the dumb server. Would this be something we'd be interested in? It could help achieve rsync-like efficiencies across filesystems and assist with downloads (but not uploads) for object storage. Unlike rsync, it would require a pre-computed file being present which makes it less useful in general.
It's impossible to do a real rsync command with cloud storage b/c you can't replace pieces of files. However, you can check for whether a file is likely to be the same and not transfer it.
There's also the concept from
zsync
which performs all rsync computation on the client side and stores it in a metadata file on the dumb server. Would this be something we'd be interested in? It could help achieve rsync-like efficiencies across filesystems and assist with downloads (but not uploads) for object storage. Unlike rsync, it would require a pre-computed file being present which makes it less useful in general.http://zsync.moria.org.uk/paper/ch02.html