An idea about how to use zk-vms with the Edge Computing Achitectures in Ethereum [cross-posted]

General
#1

This is a cross-posted article across ethresear.ch and zkresear.ch for better moderation of zk researches

@CPerezz @oskarth, Violet, and Chiro had a presentation at the zk residency group for about the future verifiable computation researches, and mentioned that “WASM does not have a gas model, so the computation is unbounded.”

So I’m just bringing up an idea here about how to use the gas model in the verifiable computational schemes.

This is a slide that I’m using that why we should contribute to the Ethereum scaling

Untitled drawing
Untitled drawing1493×282 101 KB

So one of the future scaling is something like the end-clients compute somthing on their devices and submiting a proof instead of letting other execution layer nodes to run all the computations.

In this scenario, we can have an Edge Computing Interface such as

execution(
	function,
	input vars,
	state refs,
	proof
) -> output vars: {\[key\]: value}

Then, the execution layer nodes updates the state by the output vars, and we can make the gas cost just depends on the length of the output vars.

Just a quick idea sharing :slight_smile:

6 Likes
#2

So IIUC here, the purposed idea is based on charging execution cost based on the length of the public Input/Output used by the folds done at each execution step by the VM.

Is that the case??
That sounds interesting. And it unlocks a cool idea which is how to price ZK Proofs computation based on the actual computations done.

#3

Hey @CPerezz right it’s that the case. Mostly I thought about the public outputs rather than the inputs but pricing public inputs also makes sense too!

1 Like