NEO dBFT discussion

[ssgtatn]

From the document of NEO (http://docs.neo.org/en-us/node/whitepaper.html), the dBFT consensus plays a vital role in keeping the network stable and safe. As long as the malicious nodes are less than 1/3 of the total nodes (or more exact - more than 2/3 of Consensus Nodes are honest), the network still be safe.

One question is, during the communication between the Consensus Nodes Committee (Starting from the aka "Speaker"), how many round of Messages will be sent in the Committee? In paper I thought they imply 2 rounds. Because after 1 round there might be the problem of equal values that one Consensus Node received from other Nodes (described in http://marknelson.us/2007/07/23/byzantine/)

* A round of Message is defined as the completion of broadcasting messages to all Consensus Nodes.

[toppo191]

in this paper https://people.eecs.berkeley.edu/~luca/cs174/byzantine.pdf
- n -> total number Consensus Nodes.
- m -> (n - 1) / 3, is stands for the maximum number of erroneous nodes allowed in the system

=> maximum message send in system is : (n - 1) * (n - 2) * ... * (n - m - 1)

m = 0 => n - 1 messages
m = 1 => (n -1) * (n - 2) messages

=> m is number round of Messages will be sent in the Committee

[ssgtatn]

That is not a clear answer. So how many rounds of communication will happen in the Commitee?