Frauncles; fraud proofs in REMASC

Merge-mining
1

I was thinking the other day about ZK light clients vs fraud proof light clients, because I wanted to judge whether work like https://hazync.org/ is actually worth it. And I convinced myself that fraud proof light clients are good enough, especially when one leverage ideas like 1) requiring a recent chain tip, and 2) requiring more confirmations when the block timestamps seem to suggest a chain split, even if you don’t see the split.

But then I wondered; if REMASC incentivizes miners to publish uncles, why can’t we do the same for fraud proofs?

The idea, half baked as it is;

  1. Frauncles are defined as a list of pointers to blocks that don’t belong to the canonical chain, and are alleged fraudulent, OR any block header that seems to be building on top of a Frauncle.
  2. To prove a Frauncle (at least the first one in a fraudulent fork) you need to include its header and as many transactions as necessary to prove the fraud.
  3. To incentivize including Frauncles, which is probably more expensive than Uncles, we can do two things, and probably we would need to do both:
    a) REMASC should pay miners for including Frauncles, just like we pay them for including Uncles.
    b) REMASC should pay miners more if they reuse their address between blocks, so that we have something to slash, and then the publisher of a Frauncle should get the rewards that is slashed from the fraudulent miner.

The main problems with this proposals are increased complexity, but also incentivizing pools over solo miners, but this might be a larger discussion … having something to slash from miners is useful for many usecases not just fraud proof publishing, for example DA systems.

But then again, anyone can make a smart contract that pays anyone for including a fraud proof for a parallel fork, and miners may choose to snipe that reward.

Maybe it is better to expect nothing from miners other than blindly pile work on top of work, because that is what they seem to be doing anyways.

But it appears to me that the most logical entity to create such contract and pay its rewards, is Rootstock Labs and even better; a DAO that Rootstock Labs transfer the 20% mining rewards to.

1 Like