Consensus 28/32 or 8/8?

This is a facet of XOR as soon as a node is not a node it is replaced immediately. This is the poke your finger in the sea approach I mentioned.

If a group can provide a majority decision but not a consensus decision, then the group can be automatically wider, by sending messages to each nodes close group with a signed authority stating they did the wrong thing.

This way a majority is enough to report, but not enough to make a decision and you have a gap of majority but not consensus, this is an area of reporting, but not acting on the original message, just removing bad nodes.

This is exactly the kind of rule the security sprint will put in place after feature complete :wink: There is a lot more though. We also have message direction and checking a node is allowed mathematically to have forwarded a message. This adds a lot more to the math (direction is more triangular than linear, see kademlia paper for more on this but pretty hidden like old nodes poisoning the network (which we don’t suffer from))), but is very powerful. I promise I will write some of this up, but at the moment I am with the guys working on NAT traversal and connection management in routing. So I will get back up there soon, then one day on vault logic and computation, then it gets very powerful.,

Anyway I am being a noise in this thread (sorry about that), I should let it continue as is, but getting a theoretical model and then running simulations against the xor network with the rules we have in place will be amazing.

4 Likes