Post-quantum solutions for Serai

[kayabaNerve]

The Serai stack's message-server authenticates messages with signatures. This would have to be replaced with a PQ signature scheme.

The processor/coordinator don't require any PKE other than of secret shares during the DKG protocol. That is left out-of-scope to to this issue, as it has to do with the DKGs of whichever signing schemes we adopt, which will be integration-specific (#635, #636, #637).

The coordinator does use signatures as part of its consensus protocol. We can solve this by adopting a PQ signature scheme, as our Tendermint machine is agnostic to the signing protocol, or we can replace the consensus protocol with one which doesn't rely on traditional signatures (as seen in https://eprint.iacr.org/2024/677). Relevance to #333.

The Serai blockchain itself solely uses Ristretto signatures at this time and a VRF within BABE. We'd have to replace this with a PQ signature scheme and a PQ VRF, or we'd also have to upgrade Serai's consensus protocol to an asynchronous one.

[kayabaNerve]

For Serai itself, FALCON, Rainbow are encumbered by patents with a license for any version NIST standardizes. I believe Dilithium is free of patents, even though its sibling-KEM Kyber isn't (though NIST has secured licenses for any version NIST standardizes).

[kayabaNerve]

https://eprint.iacr.org/2023/303 may be quite efficient.

[kayabaNerve]

#636 (comment) goes over efficient threshold signatures, as we need to oraclize onto Serai.