Fast Distributed Agreement

We describe a Byzantine Agreement algorithm, with early stopping, for systems with arbitrary process failures. The algorithm presented is simpler and more efficient than those previously known. It was derived using a broadcast primitive that provides properties of message authentication and thus restricts the disruptive behavior of faulty processes. This primitive is a general tool for deriving fault-tolerant algorithms in the presence of arbitrary failures.

[1]  Micah Beck,et al.  Implementation Issues in Clock Synchronization , 1986, Fault-Tolerant Distributed Computing.