分布式系统理论

FLP

One famous theory in distributed computing, known as FLP after the authors Fischer, Lynch, and Patterson, proved that in a distributed system with asynchronous communication and process crashes, processes may not always agree on the one bit of configuration.

 

CAP

 

CAP, which stands for Consistency,Availability, and Partition-tolerance, says that when designing a distributed system we may want all three of those properties, but that no system can handle all three.

 

Paxos

 

Virtual Synchrony

Byzantine Faults/Failures  - 拜占庭将军问题

Byzantine faults are faults that may cause a component to behave in some arbitrary (and often unanticipated) way. Such a faulty component might, for example, corrupt application state or even behave maliciously. Systems that are built under the assumption that these faults can occur require a higher degree of replication and the use of security primitives. Although we acknowledge that there have been significant advances in the development of techniques to tolerate Byzantine faults in the academic literature, we haven’t felt the need to adopt such techniques in ZooKeeper, and consequently we have avoided the additional complexity in the code base