FLP Impossibility

Due to the Two Generals’ Problem, we have established that we can't guarantee consensus if our network can drop messages, even under the friendly settings of the synchronous computational model. Due to FL82, we also saw that consensus is possible under a synchronous model when the network is non-faulty, but nodes can show crash-stop failures. The next question is: Is consensus possible in an asynchronous setting where a network is assumed to be non-faulty but nodes can exhibit crash stop failures? Unfortunately, it’s not possible due to a result called FLPThe acronym FLP comes from the last name initials of the three authors of this impossibility result—Impossibility of Distributed Consensus With One Faulty Process: Michael Fisher, Nancy Lynch, and Michael Paterson. This same result is also referred to as FLP85 in the research community. impossibility. We’ll understand what the FLP impossibility is in the rest of this lesson.