#Off-Chain Verifiable Proof Of State Of The Mina Blockchain

10 messages · Page 1 of 1 (latest)

woven jasper
#

Hello guys,

I'm currently investigating how Mina's blockchain sate can be verified for validity off-chain, or at least without an active connection to the internet or RCP endpoint.

That proof can come in various forms, such as:

  1. Proof that a particular transaction was submitted and acceptedon chain, what time/block that transaction was performed and, and what are the public outputs/account updates.

  2. A proof generated by a smart contract that a certain action has been performed, and included on-chain

  3. A proof of an included transaction in the history (e.g. past included blocks).

Ideally, it would be sufficient for the verifier to have knowledge of a historical state/snapshot at a certain time in the past, and be able to use that knowledge to assert the validity of future transactions/state changes.

All of this logic could be isolated in the context of a particular smart contract, and have the verification key of that smart contract assumed known knowledge. In this sense, the Mina blockchain would be used as a trusted source for application code storage, and it's associated functions.

One way of doing this would be to perhaps submit a transaction, wait for X blocks of confirmation, and then provide a historical proof of includsion of a previous transaction.

For a practial way, you can think of a "proof of deposit". For example, "here is a nullifer that allows you to withdraw 10 MINA from a particular smart contract once a proof with a particular ID is submitted into this smart contract".

How such a feat can be achieved in practice, and what limitations/hard requirements should be kept in mind?

Thank you in advance.

trim cliff
# 
Ideally, it would be sufficient for the verifier to have knowledge of a historical state/snapshot at a certain time in the past, and be able to use that knowledge to assert the validity of future transactions/state changes.

I think it's the other way around. the verifiers can assert the validity of past transactions given the root hash at a certain point in time.
So the validator needs to know the latest state of the chain or update it frequently.

woven jasper
#

That is true! I was seeking some sort of a solution were partial knowledge of the past would allow you to make assertions about near future. Since all smart contracts are verified by a single circuit that lives on the Mina nodes, I was wondering if such a setup would be possible.

But let's simplify it: let's say I have the knowledge of the blockchain state until block 100. How can a proof of state/transaction/previous block be verified based on that? From an immediate outlook, this would involve providing a valid path/witness in the chain of transactions. I'm wondering if there is a more compressed way of doing it. For example, if I know the block N of the blockchain, how can I generate a proof of the state for block N-M, with 0 >= M <= N (M larger than 0 and smaller or equal to N), and have that be verifiable offline/with partial blockchain information.

Thank you!

fathom crater
#

@woven jasper easiest technique is using write only state trees and the latest state, thats the technique we use for state proofs in Protokit. I believe the same is possible with the L1 protocol and snarked ledger hash preconditions

mellow stump
#

What are you building @woven jasper ? A bridge ?

woven jasper
#

@fathom crater from what I understand:

Write only state tree = Append-only Merkle tree , where existing entries cannot be mutated?

Latest state = most recent known block state proof on Mina?

How exactly can those two be used to verify the validity of a snapshot Mina's state off-chain? Wouldn’t that require you to have the full list of blocks in that append-only tree?

#

@mellow stump I would like to move computation over Mina state off-chain. For example, I would like to be able to verify a proof that a certain amount of $MINA has been deposited into a smart contract. This would allow for a creation of a “side-chain” which can perform computations without having an active connection to the internet

fathom crater
#

@woven jasper we solve this exact problem for Protokit L2 app chains by implementing a commitment based bridge between the L1 and the L2.

You deposit on the L1, creating a deposit commitment. Other users do the same, and these commitments come together in a hashlist.

When settling the L2, it must honor all the deposits. Otherwise you wont be able to settle

woven jasper
#

@fathom crater thank you. I’m currently looking into implementing my goal in Protokit. I’ll get back to you

mellow stump