4 ms·
Sorry, but your post came off as blatant advertising. There is no need to link to your company announcement just because it benefits you.
by gr4vityWall 2y ago
Sorry, but your post came off as blatant advertising. There is no need to link to your company announcement just because it benefits you.
- jchanimal 2y agoThanks y'all -- feedback taken. If I were saying it again I'd say something like: Merkle proofs are rad b/c they build causal consistency into the protocol. But there are lots of ways to find agreement about the latest operation in distributed systems. I've built an engine using deterministic merge -- if anyone wants to help with lowest common ancestor algorithms it's all Apache/MIT. While deterministic merge with an immutable storage medium is compelling, it doesn't solve the finality problem -- when is an offline peer too out-of-date to reconcile? This mirrors the transaction problem -- we all need to agree. This brings the question I'm curious about to the forefront: can a Merkle CRDT use a Calvin/Raft-like agreement protocol to provide strong finality guarantees and the ability to commit snapshots globally? Apologies for the noise.