🥇 Our recent work “HotStuff-1: Linear Consensus with One-Phase Speculation” reduces latency:
- 🥈 by two network-hops over [HotStuff-2],
- 🆒 thus, by four network-hops over [HotStuff],
- 🏆 while maintaining linear communication complexity against faults.
HotStuff-1 achieves latency reduction by sending clients early responses speculatively, after one phase of the protocol. Unlike previous speculation regimes, the fast path of HotStuff-1 is fault-tolerant and does not rely on optimism.
Additionally, HotStuff-1 incorporates an incentive-compatible leader rotation regime that motivates leaders to commit consensus decisions promptly.
w/ Dakai Kang, Suyash Gupta, Mohammad Sadoghi.
Read also our 📭 explaining “The Prefix Speculation Dilemma in BFT Consensus” (coming soon).
–
Stay tuned for more coming out of the [Foundations of Financial Technology (“FifTech”)] at UC Santa Barbara.
