.png){kind=icon}
.png){kind=icon}
Return
Short-HotStuff: Two-Chain BFT Algorithm With Linear Complexity and Optimistic Responsiveness
.png){kind=icon}
PRE
AI
DOI：10.1109/TIFS.2025.3595730.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
HotStuff is a pipelined Byzantine Fault Tolerance (BFT) algorithm that has the good properties of both linear communication complexity and optimistic responsiveness. However, its three-chain model has poor robustness against the performance attacks, particularly the forking attack. Some recent works have employed the two-chain model to enhance the robustness, but they either lose the optimistic responsiveness or introduce huge latency. In our work, we present Short-HotStuff, a two-chain BFT algorithm with linear communication complexity and optimistic responsiveness. In the two-chain model without waiting for a maximum network delay, the leader has to convince the other nodes that its proposed block is on a safe path when the previous leader fails, otherwise, the hidden lock will cause the liveness issue. Contrary to some existing BFT algorithms that rely on broadcasting multiple Quorum Certificates (QCs) or introduce extra phases, our Short-HotStuff requires the leader to broadcast the highest QC, as well as the evidence performed by a threshold signature with only a constant size, and such evidence can prove that its proposed block is indeed extended from a safe path. We have conducted the experiments to compare our Short-HotStuff with HotStuff, Marlin and Fast-HotStuff. The results indicate that compared with HotStuff, Short-HotStuff achieves a 35% reduction in average latency without the leader failing, and doubles the average throughput in the face of forking attacks. In the event of the leader failing, the average latency of Short-HotStuff is reduced by 43% and 82% respectively compared with Marlin and Fast-HotStuff.
Keywords:
BFT
linear communication complexity
optimistic responsiveness
forking attack
Journal
IF:
8
Papers:
5.2K
Citations:
2.3W