Closed HCastano closed 2 weeks ago
Now that we have a version number associated with each user (https://github.com/entropyxyz/entropy-core/pull/658), this is much easier as we can increase the number of subgroups for new users without needing to migrate existing users. Although decreasing it would be pretty complicated.
Having said that, as i have said elsewhere i think increasing the number of subgroups is not a good solution to the problem of increased keyshare redundancy as the network grows. I think our subgroup system needs rethinking.
We don't have subgroups anymore.
Right now we need to specify the number of subgroups we want as part of genesis. If later on in the life of the network we decide that we want to change this number, for example to increase it as the size of the validator set increases, there is no way for us to do this without restarting the network.
A couple of tricky things here:
We may be able to get away with doing some of the work in session
N-1
, and only applying the subgroup change in sessionN
.It's not reasonable to expect users to come online at some arbitrary session to re-key, so how would we get them to migrate?