Enhancing Byzantine Fault Tolerance: Leaderless and Asynchronous Consensus with Trusted Hardware Assistance

This thesis addresses challenges in Byzantine fault-tolerant (BFT) consensus protocols, proposing improvements to resilience, scalability, and performance. A leaderless BFT protocol is introduced using trusted hardware to eliminate the leader bottleneck, reduce communication complexity, and enhance fault tolerance. We propose a trusted protocol Janus that achieves linear message complexity in asynchronous networks. Juno, a hardware-independent protocol, optimizes performance without relying on trusted hardware, using a novel vector consensus approach. Theoretical analysis and experiments show these protocols advance BFT consensus efficiency, with implications for decentralized systems like blockchain, offering both hardware-assisted and independent solutions for fault-tolerant networks.