Can the Nano network undergo a contentious chain split fork?

This has been discussed on other forums but it would be useful to have documented on this forum from those who can explain it better than me.

My understanding, in summary, is that a PoW like contentious chainsplit (hardfork) is not something that is easily undertaken in Nano, if it’s even possible at all, due to the asynchronous nature of the ledger.

There have been node forks with redistribution/airdrops to existing accounts like Banano and NOS, but no chain splits.

1 Like

Since there isn't any mining you can't have race condition type forks like BTC has had where the longest chain wins. However you could get a network level split where some nodes start operating on a new version of the nano node and others stay on an old version or even go in a different direction. However this would result in significant problems with consensus on the portion of the network that can't reach >50% quorum and the threshold for quorum would need to be changed in order to allow users on the smaller network to re-delegate their voting weight. It's extremely messy and hard to accomplish compared to a BTC level fork where anyone could start mining on a new chain as it would require existing node operators to decide which version they want to run.

There are potentially cases like this already where older versions (eg before V17) that have been de-peered in the node. If enough older versions are still running they could in theory be still trying to communicate with each other, but they wouldn't be able to achieve consensus on blocks without modifications to the node since the majority principle representatives have upgraded their nodes.

6 Likes

Brainstorming some, but wouldn't it technically be possible if some percentage of network participants all decided to run their own custom node software that validated transactions in a different way? The new software might have to hardcode new representatives temporarily (or create their own form of epoch blocks), but I don't see how that would stop a fork from being created with mostly the same base DAG tree. It wouldn't affect the main Nano network since hashes would change, but you'd have two Nano-like networks running in parallel

2 Likes

It’s this extreme edge case of a new network with a resultant reduced consensus weight on the new network that I was looking for someone to explain.
Thanks Srayman, once again.