基于信任评估模型的PBFT共识算法  

PBFT Consensus Algorithm Based on Trust Evaluation Model

作　　者：任玺羽 童向荣 张伟 REN Xiyu;TONG Xiangrong;ZHANG Wei(School of Computer and Control Engineering,Yantai University,Yantai 264005,China;Secondary College,Shandong Vocational and Technical University of International Studies,Rizhao 276826,China)

机构地区：[1]烟台大学计算机与控制工程学院,山东烟台264005 [2]山东外国语职业技术大学二级学院,山东日照276826

出　　处：《山西大学学报（自然科学版）》2023年第1期108-118,共11页Journal of Shanxi University(Natural Science Edition)

基　　金：国家自然科学基金(62072392,61972360)。

摘　　要：实用拜占庭容错(Practical Byzantine Fault Tolerance,PBFT)共识算法是确保区块链等分布式系统达成一致性的重要算法,但其仍难以应用到节点数量规模较大的网络环境,且共识过程中主节点的错误选取会导致系统通信资源的浪费。针对以上问题,文章提出基于信任评估模型的PBFT共识算法(Trust-based Practical Byzantine Fault Tolerance,T-PBFT),首先采用聚类的方法对系统中的共识节点进行分组共识,降低通信复杂度;然后根据网络中节点历史行为进行信任度评估,选取可靠的节点作为共识小组中的主节点;组内实现共识后,再进行组间共识,确保系统中共识节点存储的数据信息达成一致。所提出的改进算法能够更好地应用于大规模区块链网络系统,并能够选择更可靠的主节点。通过仿真实验和分析得出,改进算法通信复杂度小于原算法的O(N2),系统的吞吐量与原算法相比最高增加了17%,共识时间节省了25%。The Practical Byzantine Fault Tolerant(PBFT) consensus algorithm is vital to ensure the consistency of distributed systems such as blockchain. However, it is still difficult to apply to a network environment with many nodes. The wrong selection of master nodes in the consensus process will lead to the waste of system communication resources. To solve the above problems, this paper proposes a Trust-based Practical Byzantine Fault Tolerance(T-PBFT) algorithm based on trust evaluation model. Firstly, the consensus nodes in the system are grouped by clustering method to reduce the communication complexity. Then, the trust degree is evaluated according to the historical behavior of nodes in the network, and the reliable nodes are selected as the main nodes in the consensus group. After the consensus is achieved within the group, the consensus among the groups was conducted to ensure that the data information stored by the consensus nodes in the system is consistent. The proposed improved algorithm can be better applied to large-scale blockchain network systems and can select more reliable master nodes. The simulation experiments show that the communication complexity of the improved consensus algorithm is less than the O(N~2) of the original algorithm, the throughput of the system is increased by 17%, and the consensus time is saved by 25% compared with the original algorithm.

关 键 词：区块链 共识算法 实用拜占庭容错 节点可扩展 信任度评估 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]