Experimental Evaluation of Asynchronous Binary Byzantine Consensus Algorithms with $t < n/3$ and $O(n^2)$ Messages and $O(1)$ Round Expected Termination

Abstract: This work performs an experimental evaluation of four asynchronous binary Byzantine consensus algorithms [11,16,18] in various configurations. In addition to being asynchronous these algorithms run in rounds, tolerate up to one third of faulty nodes, use $O(n^2)$ messages, and use randomized common coins to terminate in an expected constant number of rounds. Each of the four algorithms have different requirements for the random coin, for the number of messages needed per round, whether or not cryptographic signatures are needed, among other details. Two different non-interactive threshold common coin implementations are tested, one using threshold signatures, and one based on the Diffe-Hellman problem using validity proofs [11]. Experiments are run in single data center and geo-distributed configurations with between $4$ and $48$ nodes. Various simple faulty behaviors are tested. As no algorithm performs best in all experimental conditions, two new algorithms introduced that simply combine properties of the existing algorithms with the goal of having good performance in the majority of experimental settings.