As I understand it it’s technically not possible for validators to ignore an NNS decision and hard-fork continue running the chain as if the vote never happened / went the other way.
But if validators wanted to do that, would they have to start the network from scratch (block 0) or could they start from some checkpoint, and how far back would that checkpoint have to be?
$ICP has node providers, very different from validators in the context of traditional Blockchains. They don’t have any control over the network. The NNS does. I don’t believe there is anything a provider can do to stop an update besides disconnecting from the network.
Yes, I know, but node providers could just start over, just like they did on day 1.
The question is whether the only way for them to disobey is by starting again from block 0 and erasing all state, or if there is another way that would allow them to start from a more recent checkpoint or in some other way, and what that would look like.
Trying to better understand the security model of the IC and the assets it hosts.
And why would you need to do a hard fork?
To my understanding you can’t. But it would be fun to watch and see how much money and resources is wasted trying to do a hard fork best of luck!
Not sure what you mean but I certainly don’t “need to do a hard fork”.
More broadly, if the NNS votes to eliminate your personal balance of ICP, you may be interested in knowing your available options.
And more broadly still, understanding the security model of the IC, including how it might deal with attempted hard forks, helps us build more secure applications.
Isn’t every network upgrade a “hard fork”?
I think this was part of @yvonneanne’s presentation during the last Global R&D meeting.
Maybe technically but I wonder what would happen if there is a strong disagreement between NNS and node providers.
Basically, I’m trying to better understand the censorship-resistant properties of the network.
Which is different from a Hostile Takeover