seems do-able with those two pieces of tech
@requiem I think there are a couple. Tahoe project may be one: https://tahoe-lafs.org/~warner/pycon-tahoe.html
... I'm having a mental blank on the name of another similar one!
@requiem ah! the other thing I was thinking of is IPFS - https://ipfs.io/ .
I think there is a project built on top of IPFS that does something like this... although I don't remember what it's called, sorry
@requiem definitely get a "really powerful, also really complex!" energy from it
@kemonine @requiem I was thinking the other day that it'd be great if I could add an untrusted syncthing device to my network that synced ciphertext of some things I already sync, to increase the overall availability - but that device couldn't acess the plaintext of what it was syncing.
(Haven't taken the time to look at the SyncThing protocol to see if this is even remotely possible, this is at the "idle thought" stage)
Syncthing shovels filesystem data about without much thought or care about what that data is
It has a conflict resolution that's as robust as "these two things tried to collide, pick a version thats proper" and that's about it
You'd need something like encfs or similar 'on top' of whatever syncthing is syncing to get that kind of support
it's likely do-able but more work than doing the crypto at the block layer and leaving syncthing's transport security to do the hard part of moving data around safely w/o crypto actually applying at the fileystem level
@requiem
Also: last I looked nextcloud has some options for object storage, encryption and the like, it may actually be easier to build a nextcloud cluster than other alternatives
@requiem
@requiem this used to be a main feature of CrashPlan but they removed this and made Crashplan into a corporate-only tool about a year ago! http://www.documentsnap.com/how-i-do-offsite-backup-to-a-friends-computer-using-crashplan/
@requiem I want something like this,
for my general archives. no blockchain needed, just common sense and a few mutually trusting people.
Everybody runs a store, each store lazily checkpoints its blob archives to the others.
- blob store archive: haystack https://research.fb.com/wp-content/uploads/2016/11/finding-a-needle-in-haystack-facebook-s-photo-storage.pdf
- data integrity: blake3 fast hashing https://github.com/BLAKE3-team/BLAKE3
- lookups: xor filter https://lemire.me/blog/2019/12/19/xor-filters-faster-and-smaller-than-bloom-filters/
- automatic distribution: TBD
- retrieval: just use nginx + http range query
https://tahoe-lafs.org/trac/tahoe-lafs
@requiem Sounds like Tahoe-LAFS is the software you're looking for. https://tahoe-lafs.org/trac/tahoe-lafs
- replies
- 1
- announces
- 0
- likes
- 1
@kazriko @projectgus @kemonine @requiem
🤔 Does that work with n:m source:destinations?