An example of BFT is Bitcoin, a peer-to-peer digital cash system.  The Bitcoin network is working in parallel with the creation of a proof-of-work blockchain that allows the system to overcome Byzantine failures and obtain a coherent global view of the state of the system. The problem is complicated by the presence of insidious generals, who vote not only in favour of a suboptimal strategy, but also selectively. For example, if nine generals vote, four of whom support an attack, while four others are in favour of withdrawal, the ninth general may send a withdrawal vote to these generals in favour of withdrawal and one vote to attack the rest. Those who have obtained a withdrawal vote from the ninth general will withdraw while the rest will attack (which may not be good for the attackers). The problem is further complicated by the fact that generals must be physically separated and send their votes through messengers who might not vote or falsify false votes. Byzantine errors are considered the most common and difficult class of errors among error modes. The Fail-Stop-Fail mode takes the simplest end of the spectrum. While fail-stop error mode simply means that the only way to reach the defect is a node crash detected by other nodes, Byzantine errors do not involve constraints, meaning that the undone node can generate any data, including data that make it appear as a functional node. Thus, Byzantine errors can confuse error detection systems, making the margin of error more difficult. Despite the analogy, a Byzantine failure is not necessarily a security problem with hostile human interventions: it can be the result of electrical or software errors.
The system resists Byzantine errors when a component A -Displaystyle A can send a value x -Displaystyle x,” then: A Byzantine error is an error that presents different symptoms to different observers.  A Byzantine failure is the loss of a system service due to a Byzantine error in systems that require consensus.  Several solutions were described in 1982 by Lamport, Shostak and Pease.  They began by saying that the generals` problem can be reduced to solving a “commander and lieutenant” problem, in which loyal lieutenants must act together and that their action must correspond to what the commander ordered in the event that the commander was loyal: the typical assignment of this story to computer systems is that computers are generals and that their links to digital communications systems are Bots.
（转载本站文章请注明作者和出处 酷 壳 – CoolShell ，请勿用于任何商业用途）