Scalable Byzantine Consensus

Consensus, state-machine replication (SMR) and total order broadcast (TOB) protocols lie at the core of most blockchain systems and many other distributed applications. However, they are notorious for their poor scalability. Despite the recent race in reducing overall message complexity of leader-driven SMR/TOB protocols, scalability remains poor and the throughput is typically inversely proportional to the number of nodes. In this session we will hear from Matej Pavlovic regarding MirBFT and its successor, ISS, turn leader-driven protocols into scalable multi-leader ones. The design is general enough to accommodate most leader-driven ordering protocols (BFT or CFT) and make them scale. The research prototype implementation, when applied to 3 different protocols (PBFT, HotStuff, and Raft) at the scale of 128 nodes, shows significant performance improvements (37x, 56x, and 55x, respectively). Now is the time to build a public, open-source, highly modular, production-ready implementation of MirBFT/ISS for use in real-world systems. Presenter Bio: Matej Pavlovic is a distributed systems researcher specializing in Byzantine fault-tolerant consensus protocols. He obtained a Master's degree in computer engineering at the Vienna University of Technology in 2013 and a PhD in distributed systems in 2019, supervised by Prof. Rachid Guerraoui. After having worked on permissioned blockchain solutions at IBM Research Europe - Zurich, he is joining Protocol Labs to work in the domain of public blockchain systems. Recently he has worked on showing that consensus is not necessary to implement a cryptocurrency (The Consensus Number of a Cryptocurrency, PODC'19) and on scaling classic BFT protocols (MirBFT, pending publication). Currently he is starting the effort to implement a highly modular production-grade open-source implementation of MirBFT and related protocols - mirbft github