概率信任驱动的分层异步BFT协议

打开文本图片集

中图分类号：TP309 文献标志码：A 文章编号： 1000-5013（2026）02-0153-11

Probabilistic Trust-Driven Hierarchical Asynchronous Byzantine Fault Tolerance Protocol

LI Shuting，GUO Rongxin，SHI Yifan （College of Engineering，Huaqiao University，Quanzhou 362o21，China）

Abstract：To address the chalenges of node heterogeneity，dynamic communication，and frequent failures in decentralized physical infrastructure networks，a probabilistic trust-driven hierarchical asynchronous Byzantine Fault Tolerance （PT-HABFT） protocol is proposed. The proposed protocol integrates a Gaussian-Bayesian reputation model that combines node geolocation，network latency，and historical behavior into multidimensional feature vectors to achieve dynamic trust evaluation and access control. An asynchronous hierarchical structure with decoupled sub-layer local consensus and upper-layer global aggregation is adopted，and a pipeline verification mechanism is designed to support multi-stage concurnt processing. Furthermore，a reputation-driven weight adjustment and view-switching strategy is developed to enhance robustness Theoretical analysis demonstrates that the proposed protocol satisfies both consistency and liveness under Byzantine fault tolerance assumptions. Experiments on the Bamboo framework show that，with 128 nodes，PT-HABFT achieves approximately eightfold increase throughput and 58% lower communication overhead compared with the HotStuff protocol.

Keywords：Byzantine Fault Tolerance;Gaussian-Bayesian reputation model; multi-source information fusion; hierarchical consensus； parallel asynchronous structure

随着去中心化物理基础设施网络的兴起，分布式共识机制正逐步从传统金融应用拓展至边缘计算、智能电网、物联网等新兴场景[1]。（剩余16363字）