Node Compensation Question

economics
#1

Hey Everyone,

I am playing around with some ideas for rewarding full nodes for the work they do, especially given worries of state growth in the future. I’m wondering what previous discussions exist regarding the feasibility of compensating full and archival nodes. Are there ways to fairly validate that nodes are actually performing work and are not malicious without resorting to proof of work or staking? Are there perverse incentives that pop up when rewarding nodes?

To keep things simple, I’m assuming that the funding method does not come from inflation, but from some other revenue stream like state rent or ENS rent.

#2

VIPNode is doing something in this area :slight_smile:

#3

Hah, this is exactly what Eric sent me a few days ago when I posted a similar question on Twitter. (Sorry I’m asking this again Eric!) That’s an interesting project but is looking at it in a way that isn’t really helpful for what I am thinking about. My understanding is they are directly compensating full nodes by creating a way to charge light nodes for slots. That feels a little clunky to me as it increases the friction to run a light node. Eventually the ratio of light nodes to full nodes would drop, lowing the overall rewards for full nodes.

What I’m trying to look at is whether new revenue streams like state fees and/or ENS rent could be directed to full nodes and distributed for the work that they do. I have grand visions of something more complicated where the revenue stream is paid into a DAO/smart contract and new tokens are minted to pay nodes based on bonding curves, with outside investors also able to buy in to the revenue stream and fund grants. But that’s a million steps away and probably over-complicated anyway. I’d like to work up from a simple system, and to do that I want to understand the economics of rewarding full nodes. hence the previous questions.

Or am I thinking about vipnode all wrong?

#4

One big challenge with this is that you’d need to know in advance what all the nodes that would ever process those txns would be. I think because of that, bundling the “node compensation” to when ether is sent may be easiest. Vitalik had proposed the 1gwei/txn idea that clients could implement. There, the challenge becomes determining at what level the fee is paid (client, wallet, or dapp?), but that seems more manageable.

1 Like
#5

I remember reading Vitalik’s proposal for a 1 gwei fee that goes to wallet development and I think that is already influencing my thinking. AFAIK his fee proposal would get paid straight from the user making the transaction to the medium of transaction (wallet). Could this be translated to nodes by having a flat fee on transactions that goes directly to the full node that broadcasts your transaction?

I have in my head an idea to pool the fees to distribute in more exotic ways, but as mentioned here and in other forums this is not very sybil resistant. Has anyone looked at a system where nodes can submit a record of the blocks that they validated (within a certain timeframe of course) and then X blocks down the road, when the chain has enough confirmations to feel safe there wont be a reorg, the submitting nodes that distributed the correct block would be paid out of the node reward pool.

(But I’m just some dude with a limited knowledge of how block propagation works so this could be totally unfeasible.)

#6

It might not be a perfect fit but have you looked into stateless clients? https://ethresear.ch/t/the-stateless-client-concept/172