Author also discussing it on their Twitter [0]. I’m reserving judgement for now, but this has lots of potential usecases even apart from replacing postgres.
I'm already quite excited about Turso being SQLite-compatible, but adding many features on top.
And when a feature is not directly compatible with SQLite (ie: you can't directly read the file with `sqlite3`, it's straightforward to convert). This is great because you know you'll always be able to continue working with that database. Even if Turso stopped working, it's still a valid SQLite database.
A combination I would be excited about is:
- Full support for Postgres protocol/wire format (ie: Postgres, but in-process, backed by a single file).
- Optional: Client/server architecture for further scaling and remote management using existing Postgres tooling
- All backed by a SQLite-compatible file
They are already adding MVCC to SQLite anyway. So their effort seems doable, and I hope they succeed.
Supporting Postgres is a good goal but honestly the real challenge is extensions. Would supporting Postgres wire compatibility guarantee any Postgres extension would also work? This is one of the problems with Aurora - it’s all well and good until you need an extension that isn’t on the blessed list
curious how hard it is to implement wire protocol compatibility vs the internal api surface. Are we talking like, an order of magnitude more work here? I wonder if, for example, Aurora team are literally maintaining forks of extensions for their implementation and that's why only a subset of all possible extensions are supported?
> If you have ever babysat a REFRESH MATERIALIZED VIEW cron job, or faked live views with a pile of triggers, read that sentence again: our views update themselves, live. This is the feature Postgres users have quietly wanted for twenty years.
Real-time materialized views - so far mostly Chinese tech shops have solved this - look at Apache Doris, StarRocks - but these are built on top of MySQL then maybe RisingWave which is Postgres wire compatible.
having real-time materialized views means you can take out 1 or 2 infra pieces from your stack e.g Flink & Kafka - for some simplified use cases. that means no more ETL jobs for some use cases.
[0]: https://x.com/glcst/status/2077759127682486561?s=46