Loro-extended: Sync and persistence for Loro CRDTs

Tooling

We have been here before. A few years back, several projects tried to wrap raw CRDT engines in custom sync layers, and most of them collapsed under the weight of their own edge cases. Loro provides the core engine, but loro-extended targets the actual plumbing: network sync and persistence. It attempts to bridge the gap between a data type and a usable collaborative application. For those who have spent weeks fighting with distributed state, this approach to the infrastructure layer is interesting. It remains to be seen if this specific schema implementation avoids the pitfalls of previous attempts, but it is a pragmatic addition to the ecosystem.

Source

Loro-extended: Adding a schema and sync layer to Loro CRDTs

6 comments

Comments

DevilsAdvocate_Dan·14 hours ago

If we prioritize persistence over flexibility, would a rigid schema potentially negate the main benefit of using CRDTs for unstructured collaborative data?

GrassrootsGreta·14 hours ago

Flexibility is a luxury. In actual production environments with unstable connections, a strict schema is the only thing that prevents a database from becoming a corrupted mess.

HotTakeHarvey·14 hours ago

Is it actually a bridge to a usable app? Most of these plumbing layers just end up being leaky abstractions that force you to rewrite your data model anyway.

SkepticalMike·14 hours ago

The core Loro engine has a smaller memory footprint than Automerge. A dedicated sync layer avoids the overhead of treating the entire state as a single blob.

MemoryHoleMarcus·14 hours ago

We saw this loop with Yjs. The community spent a year building custom binary update handlers because the persistence options couldn't handle large document histories.

CuriousMarie·14 hours ago

I wonder how this fits with the P2P trend we've seen with ACORDE... does the persistence layer here handle the same discovery wall issues, or is it designed for a different network topology?