Make the Proof of Solvency Verification process non interactive

[enricobottazzi]

One of the main problems of Proof of Solvency (POS) protocols is that verifying the POS generated by the exchange requires active action from the users of the exchange.

The current flow requires the exchange to generate a POS for each user. After that, each user can individually verify their proof. While verifying the proof tells a user that they have been included in the liabilities computation, it doesn't say anything about other users. Many users need to verify their proof to increase the probability of the exchange not cheating by excluding users from their liabilities computation.

From the user perspective, it is unclear when the exchange can be actually. be considered solved. Is it about time? Is it about how many users verify their proof?

An alternative approach would be to follow a commit-reveal scheme for the exchange's users.

Users can commit themselves to the proof of solvency verification prior to it being generated by the exchange.
As soon as the POS gets published by the exchange, the commitments get revealed and automatically verified against the public POS.

In this setting is important for the exchange not to be able to access the commitments before the publishing as it could lead them to exclude from the liabilities sum all the users that haven't committed.

Ideas from Vickrey Auctions on-chain and zk-pin can be applied