MEV-Airnode Sentiment Check

Team: MEV-Airnode

Period: 3 months

Amount: $36.9k

Destination: TBD multisig

Summary:

This proposal requests funding for the development of an auction mechanism on top of Airnode designed to monetize for MEV. This will be implemented as an open source API that data providers may run alongside Airnode. This API will host sealed bid auctions at frequent intervals to assign the right for third parties to perform Beacon updates with signed data. The proposal is written by the primary agent.

Abstract:

Oracles can extract MEV by bundling data feed updates with other state changes like liquidations into the same transaction. This service will use Airnodes signed data gateway to offer rights to update Beacons to the open market of searchers. Ensuring separation of the oracle and searcher roles helps avoid vertical integration, while regular public auctions increase competition and transparency, both helping to mitigate the centralising forces of MEV. High competition between searchers maintains the security of the dapp, while helping most of the value flow back to them as can be seen with MEV-Geth auctions trending towards 95% of available value paid to auctioneers (miners).

Motivation:

This service should increase the adoption of Beacons by lowering the costs for dapps, and increasing the revenue for data providers. Oracles serving Maker, Compound, and Aave on Ethereum mainnet alone forfeited a lower bounds estimate of $15,800,000 in MEV from May 21st to June 21st (calculations at the end), which helps show the size of the industry. It will also make Beacons more responsive to volatility by removing the necessity for deviation thresholds to be crossed to perform an update. Lastly it will act as a public good by moving gas price bidding wars off chain, reducing negative externalities of MEV.

Specification:.

  • MEV-Airnode will be an API implemented in Python that runs alongside an Airnode instance. Searchers can submit sealed bids to this API along with the associated request parameters for the respective Beacon, and at regular intervals it will assign a winner of the auction to the highest bidder. It will then make an HTTP request to the Airnodes signed data gateway with the parameters provided from the winning bid, and re-sign the values returned in the signature along with the bid amount and the wallet address of the bidder.
  • Only the winning bidder will be allowed to view the API response for that auction round. So that MEV-Airnode access will not act as a substitute to API access for the underlying data provider, and searchers will have to separately purchase that to participate effectively.
  • This response can then be used by searchers to update the beacon using a function called updateBeaconWithSignedBid() that will verify that the msg.sender of the tx is the winning bidder, and the value paid to the data provider is higher than the bid amount signed for (as well as the other functionality of the updateBeaconWithSignedData() function).
  • To prevent denial of service attacks MEV-Airnode will not allow searchers to win two auctions in a row, and will revoke api-keys of searchers who propose winning bids and consistently fail to execute the corresponding Beacon update.

Drawbacks:

  • Searchers who access MEV-Airnode could use it to have first rights to MEV on every protocol that uses that beacon. We could get around this if we had to by having a provider deploy a second beacon instance using a second wallet specifically for the protocol that wants to take part.
  • Added complexity and extra software for data providers to run. These risks can be mitigated with a slow rollout across providers and beacons.

Deliverables:

  • MEV-Airnode.
  • PR on DapiServer.sol to add updateBeaconWithSignedBid function.
  • Articles and documentation on the service.
  • Data collection/monitoring software for the service.

Timeline:

  • 3 months

Requested Funds:

  • Product Lead: doge_bull, salary: 4.3k per month
  • Lead Developer: Elemental_Brian, salary: 8k per month

Calculations:

  • These are a bit rough but should serve as a decent lower bounds.
  • Source: Dune on 06/21/22.
  • I took total debt covered in the last month and multiplied that by 5% (lower bounds liquidation incentive) to get the total profits extracted. This number assumes searchers do not need to bribe block producers as they cannot be frontrun due to the data being signed for only them to update the Beacon with.