communication complexity
You can find that the number of network messages drops from O(N N) to O(N), but because each prepareQC contains 2f + 1 signatures, the communication complexity is still O(N N ).
Advances in cryptography: prepareQC containing 2f + 1 signature information, using threshold signatures , reduces the occupied space to O(1), and reduces the complexity of the entire communication volume to O(N)
communication model
The communication model defines the power limit of the adversary to delay the message. There are three basic communication models: synchronous model, asynchronous model and partially synchronous model.
-
Synchronization model: Assuming the maximum network delay T, it must be set relatively large (min level?). Because of the actual situation, there may be problems in the network, and it may take a few minutes, or a few hours, or even a few days to solve the problem.
-
Asynchronous model: There is no guarantee of network delay, and the network cannot guarantee consensus. The famous FLP theorem
-semi-asynchronous model: partial synchronous model. The compromise between the above two models is also called partial synchronization. The network may be in an asynchronous state, but GST (global stable time) will enter the synchronization state.
Most of the consensus algorithms are based on the semi-synchronous model, which is also in line with the actual situation. You can refer to the following.
https://decentralizedthoughts.github.io/2019-06-01-2019-5-31-models/
Responsiveness
Once the network enters a synchronized state, a good leader reaches a consensus at the speed of actual network delay.
Also called (Optimistic) Responsiveness Optimistic responsiveness, I think it is more intuitive. That is, after the network reaches the synchronization state (GST), that is, after all network problems are solved, the time required to actually finish the work is consistent with the actual network delay. It is similar to moving 1,000 bricks. Now that the bricks are ready and the people are full, then assuming that 10 bricks are moved in one minute, they can be moved in 100 minutes.
(Optimistic) Responsiveness After GST, any correct leader, once designated, needs to wait just for the first (n−f) responses to guarantee that it can create a proposal that will make progress. “As fast as the network propagates, on a good day”
safety
Bad things will never happen, that is, it is impossible for all good nodes to submit (commit) conflicting data.
Pay attention to distinguishing the security of cryptography. This concept has nothing to do with the consensus algorithm. The security in the blockchain is guaranteed by cryptography.
activity
As long as the system network enters a synchronized state, the system can eventually reach a new consensus.
In the blockchain, the height of the chain will increase, and new blocks will continue to be produced.
Reference link https://blog.csdn.net/wenyuanhust/article/details/125495449