A survey of distributed consensus, two phase commit, etc papers over the last 30 years

Mathematical Foundations of Computer Science, 2000

In a classic paper, Garcia-Molina specifies the leader election problem for synchronous and asynchronous distributed systems with crash and link failures and gives an elegant algorithm for each type of system. This paper points out a flaw in Garcia-Molina

Several papers that may be of use

We present the first protocol that reaches asynchronous Byzantine consensus in two communication steps in the common case. We prove that our protocol is optimal in terms of both number of communication steps, and number of processes for two-step consensus

Includes a specification of the protocol in TLA+ which may prove useful.

This project is aimed at developing algorithms and implementation techniques to build practical Byzantine-fault-tolerant systems, that is, systems that work correctly even when some components are faulty and exhibit arbitrary behavior.

Implementation of the Paxos at War protocol in Java.

The optimistic latency of Byzantine Paxos can be reduced from three communication steps to two, without using public-key cryptography.

Lamport's Paxos