WASI Level 2 - Are we nearly ready for a WASM SDK?

Show & Tell
1

Firstly, Temporal is fantastic. Utterly fantastic. I feel like I’ve 100x’d since making the jump.

This is my first post in the community forum, so I apologise in advance if this is the wrong place to discuss this, but the idea has been bouncing around in my head for a few days now.

Now that WASI is maturing and we have features like wasi-clocks, wasi-sockets and wasi-random, it feels like there could be scope for a Temporal WASI / WASM SDK. I’m quite excited by this, as it could give us scope to run conventional binaries as part of a deterministic workflow, along with a whole host of other possibilities (deterministic PRNGs, filesystem interception anyone?).

The Temporal sdk-core docs make a passing mention of bindings to a WASM interface. Is this something that is already in active development? Or, if others think this is an interesting area to explore, I’m toying with the idea of putting a wasmtime based rust prototype together.

I may be overlooking a reason why this is a bad idea, but would love to hear any thoughts on the topic, postive or negative.

Thanks! Tom.

2

Tom, we are watching WASM closely. The main limitation is that it is not practical with languages requiring runtime. Each instance includes the whole runtime, so it is not possible to cache many of them. So, all the real production applicaitons that compile to WASM are written in C/C++/Rust.

1 Like
3

Hi Maxim, thanks for clarifying. I can see why that limits the scope of a WASM SDK at the moment.

Great to hear that it’s still very much on the radar though.

All the best,

Tom.