Byzantine Fault Tolerant State Machine Replication in Any Programming Language

Ethan Buchman · 2019

State machine replication is a fundamental primitive in fault tolerant distributed computing, but few production tools exist to support the replication of arbitrary state machines. The tools that do exist, like Apache Zookeeper, CoreOS's etcd, and Hashicrop's Consul, include an implementation of a consensus algorithm (eg. ZAB or Raft) for replication, and a service-discovery oriented key-value store as the state machine. While these tools can tolerate crash failures, they cannot tolerate malicious or adversarial ("Byzantine") faults.

Read the paper · More papers on PaperTik