Closed ggutoski closed 3 years ago
I noticed that the composite dlog proof for the Paillier key pair is also being generated for every share. So, like the Zksetup we should only prove it once, and use the rng to generate it.
Based on offline discussion, we agreed that ZkSetup
, which is being randomly generated, should use the PRNG (This has been solved in #160. Furthermore, we could make it so that the proof for the Zk setup is also deterministically generated with a PRNG. This part of the issue will be tackled in #163
Currently we sample fresh ephemeral randomness to compute the
ZkSetupProof
. https://github.com/axelarnetwork/tofn/blob/d5d00ee459e3b921e33a0474831dd465ceddb7e4/src/gg20/crypto_tools/paillier/zk/mod.rs#L50 Thus, we cannot deterministically deriveZkSetupProof
from a prng seed. This isn't currently a problem because we currently have no need to deterministically re-derive zk setup info. But it's silly to use the prng only for the keypair portion of the zk setup but not theZkSetupProof
portion.