feat: experimental incentivize store protocol

alrevuelta commented 9 months ago

Context

The goal of this issue is to create a proof of concept branching out nwaku, with the following new feature. Incentivize store protocol requests requiring a proof of payment of x tokens to an address in order to be processed. After successfully proving that the payment was done, then the node will return the requested messages. Note that this is intended to stay as an experimental branch, and most likely not merged into master in the short term.

For context, see this.

Store protocol is one of waku's service protocols, based on a simple req/resp protocol where a node (full or light) requests the messages of a given topic from start/end dates and gets them in return. The implementation is here, which is triggered when a node receives the request.

Feature

Some informal requirements. Let requester be the node that requests messages from store and responder the full node that stores messages and offers store capabilities:

Introduce a new req/resp interaction that is done before the actual store req/resp (explained above). In this new interaction, both requester and responder agree on a price and amount of messages for given topic that will be paid and served. The requester gets an invoice id to be paid.
The requester pays the invoice and performs the store request including this proof. This would require modifying request.
After the responder verifies that the payment was completed, it returns the agreed messages, completing the interaction.

Some notes:

We can use Ethereum, Sepolia testnet would be fine, as we have nodes to use, and RLN is already integrated with this. Of course in reality it wouldn't be feasible to do this due to fees and processing times, but for the proof of concept its fine. As long as we target EVM (or Ethereum RPC compatible chains, we can then change).
Nodes must have a new parameter, where they configure the "fee recipient" aka address where they will be getting the fees.
To cut scope for this proof of concept, we can avoid having smart contracts and just use vanilla rawTx, using the data field to include the invoice id explained above.

The deliverable would be a branch with such feature, where multiple wakunode2 instances can be run + interconnected in requester and responder mode. It must include the instruction (cli commands, etc) on how to see this feature live, where a node pays some ETH in return of a message.

chair28980 commented 8 months ago

No deadline set as this is highly experimental task.

s-tikhomirov commented 7 months ago

An observation extracted from a discussion in the research repo.

Storing nodes are supposed to take on risk without being sure that they will get reimbursed in the future. A full node must store non-ephemeral messages as they arrive, in the hope that someone, sometime in the future will be interested in them. This is akin to PoW mining where miners expend energy first hoping to get rewarded later, so that's not to say this cannot work, but it's something I'd keep in mind.

s-tikhomirov commented 7 months ago

Weekly Update

achieved: wrote the first draft of incentivization outline
next: discuss open question, continue structuring the document

fryorcraken commented 7 months ago

@alrevuelta should we move this issue to the research repo?

alrevuelta commented 7 months ago

@fryorcraken We have a tracker for the idea here https://github.com/waku-org/research/issues/21 but I would like to keep this one in nwaku, since this issue relates to a nwaku implementation with code, implementing the idea of https://github.com/waku-org/research/issues/21

s-tikhomirov commented 7 months ago

Weekly Update

achieved: started gathering feedback around incentivization outline https://github.com/waku-org/research/pull/33
next: continue the discussion

s-tikhomirov commented 7 months ago

Weekly Update

achieved: (was mostly afk last week)
next: merge initial version of the incentivization outline; set priorities w.r.t. next steps