It would be nice to combine the best parts of restic and Borg with a mechanism for hostile environment, (and if it is connected to the Internet, then it is a hostile environment.)
I was imaging a server that the client can write new date to and read from, but can't edit previous versions without (one or more of: [master key, {multiple} 2FA ] ).
Also distributed backup seems a low-hanging fruit when each household has so much spare drive space on devices. I sketched a P2P backup that was aware of which devices where onsite and which were at a different location, (tested using an old laptop on the end of a friends ADSL line that acted as a node in my private "cloud" storage. The daemon on the laptop authenticated my anonymous identity, as a friend and controlled how much of its spare space it was willing to donate to me, (either a percentage of free space or "egalité" where it would offer as much as I was offering it.) The part that I didn't crack was getting my local daemon to shuffle n+1 copies of blocks about to maintain availability when friends ADSL would go offline or would be saturated with stream to the point where the latency was too high to be usable.
I currently use https://www.duplicati.com/ and like some of the features in https://perkeep.org/doc/compare#compare http://nongnu.org/libchop/ is probably worth mentioning, (though not a solution in and of itself.) https://syncthing.net/ Other things that have some overlap with this domain: https://github.com/gridsync/gridsync https://github.com/RetroShare https://gnunet.org/ http://oceanstore.sourceforge.net/ http://www.openafs.org/ http://www.coda.cs.cmu.edu/ http://duplicity.nongnu.org/ https://www.syncany.org/
It would be nice to combine the best parts of restic and Borg with a mechanism for hostile environment, (and if it is connected to the Internet, then it is a hostile environment.)
I was imaging a server that the client can write new date to and read from, but can't edit previous versions without (one or more of: [master key, {multiple} 2FA ] ).
Also distributed backup seems a low-hanging fruit when each household has so much spare drive space on devices. I sketched a P2P backup that was aware of which devices where onsite and which were at a different location, (tested using an old laptop on the end of a friends ADSL line that acted as a node in my private "cloud" storage. The daemon on the laptop authenticated my anonymous identity, as a friend and controlled how much of its spare space it was willing to donate to me, (either a percentage of free space or "egalité" where it would offer as much as I was offering it.) The part that I didn't crack was getting my local daemon to shuffle n+1 copies of blocks about to maintain availability when friends ADSL would go offline or would be saturated with stream to the point where the latency was too high to be usable.
Good luck.