Secure Causal Atomic Broadcast, Revisited

Abstract

We revisit the problem of preserving causality in Byzantine fault-tolerant (BFT) atomic broadcast protocols, a requirement first proposed by Reiter and Birman (TOPLAS 1994). While over the past three decades, this requirement has been met through the deployment of expensive public-key threshold cryptosystems, we propose three novel, secure causal BFT protocols without using public-key cryptography. We implement and evaluate these protocols, showing that they significantly outperform existing constructions that use threshold cryptosystems.