Assuming that the validator set rotates 1/4 each time
Q1: Do we rotate the validator set simply because it’s too large to have everyone vote on every block?
where finality could break if two consecutive validator sets sign conflicting data
Q2: If we were to sacrifice some liveness through increased block times in exchange for larger validator sets, would we not be able to have all the validators participate in every round and just ommit this problem altogether?
We have lots of signatures flying around lots of p2p subnets, and it’s pretty expensive data-wise though there are ways to optimize it.
Q3: My thoughts/experience point me to either FROST DKG or BLS signature aggregation. Wanted to know if that’s what you had in mind or something else I should research?
Disclosure: Though these comments are relevant in the context of Ethereum, my motivation is to transfer the learnings over to the research and prototyping of v2 of the protocol my team is working on. Happy to be an open book upon request but don’t want to add noise here.