sqld VS donutdb

There's https://github.com/libsql/sqld , but sqlite's concurrency model doesn't always work well with long-lived transactions (and just the network hop can be slower than a local transaction), especially if you want to write.

For instructions on how to build sqld from source, see the docs.

I also discovered https://github.com/libsql/sqld/ which provides similar functionality ("Bottomless replication"), and exposes the database over the network via the Postgres protocol. I haven't played with this yet, but it would be neat to try running some apps that require Postgres with this.

There are some language-specific clients suited for data transaction with a Turso database, such as @libsql/client for TypeScript, libsql-client for Python, and libsql-client for Rust(currently works with Cloudflare workers), but for simplicity, we’re going to use curl scripts. We’ll interact with the database by sending HTTP POST requests containing the JSON-encoded SQL queries that we’d otherwise run on the SQL shell as already demonstrated. By supporting database interaction via the HTTP protocol, Turso guarantees easy access from all types of applications, especially edge functions.

- And you can embed it in your application and it will really talk to a "SQLite" database over the network: https://github.com/libsql/sqld (not a lot on consistency, I wonder how multiple writers are handled?)

Man this is cool. While I really enjoy my own solution of using a custom SQLite vfs that stores your db transparently in dynamodb[0], this really is a compelling alternative.

I wonder how viable this would be to use from aws lambda? It seems like the way lambda does concurrency probably doesn't play all that well with litefs. Maybe it's time to move some workloads over to fly.io.

[0]: https://github.com/psanford/donutdb

There's a bunch of projects that have implemented this. I wrote a SQLite VFS in Go that lets you query a read-only SQLite db over http (including from s3) [0].

The VFS API offers the possibility for weirder storage solutions, if thats the type of thing you're into. Recently I've been moving some of my personal websites hosted on AWS Lambda over to use a read/write sqlite db backed by DynamoDB[1]. There are a bunch of limitations to this type of thing (like it uses a global write lock), but it works nicely for DBs that have low write frequency.

[0]: https://github.com/psanford/sqlite3vfshttp

[1]: https://github.com/psanford/donutdb