A Performance Comparison of Algorithms for Byzantine Agreement in Distributed Systems

Abstract

Reaching agreement in the presence of byzantine processes is an important task in distributed systems. Theoretical analysis of algorithms for Byzantine Agreement can provide insight into their efficiency. However, analysis of algorithms under varying parameters and practical constraints through experimental evaluation can be key to understanding the performance and trade-offs of theoretically well-performing algorithms. We compare the performance of two randomized byzantine agreement algorithms-one using the pull-push approach and another using the concept of quorums-and a third recent simple deterministic byzantine agreement algorithm. Through implementation on a testbed environment using the metrics of bit complexity, round complexity and latency in the presence of network sizes and faulty processes, we quantify the performance of each algorithm. In terms of bit complexity, we show that for small networks (n <; 32) and up to 10% faulty processes, the simple deterministic algorithm performs best, while for larger networks, pull-push is the best performing algorithm. The second randomized algorithm performs best in terms of latency.