首页|基于Raft的多主节点拜占庭容错共识机制

基于Raft的多主节点拜占庭容错共识机制

扫码查看
为了解决联盟链中实用拜占庭容错(PBFT)共识机制在区块链网络中节点数量增多的情况下,通信复杂度高、共识效率低下等问题,本文提出一种基于Raft的多主节点拜占庭容错共识机制IMRBFT。IMRBFT通过Maglev 一致性哈希算法对区块链网络节点均匀分组,将这个共识流程分成组外共识和组内共识2部分。组内先选出领导者节点,通过信用机制将节点分为3个等级:可信节点、普通节点和不可信节点。与投票机制共同降低恶意节点成为领导者节点的概率,并与其他组的领导者节点组成委员会,委员会再经过组外信用值机制选出信用值最高的多个主节点进行组外PBFT共识。组内共识在Raft共识的基础上引入监管节点与中继节点,进一步提升安全性与共识效率,减少恶意节点的作恶行为。实验结果表明:IMRBFT的通信开销为线性增长,通信量为PBFT的41。6%,吞吐量为PBFT的4。2倍,共识延时降低76。4%。随着节点增多,优化更加明显,完全满足大型区块链网络的通信复杂度小、吞吐量高、共识延时短、安全性与共识效率高的要求。
Multi-primary-node Byzantine Fault-Tolerant Consensus Mechanism Based on Raft
In order to solve the problems of high communication complexity and low consensus efficiency of practical Byzantine fault-tolerant(PBFT)consensus mechanism in the consortium chain under the condition of increasing number of nodes in the blockchain network,a multi-primary-node Byzantine fault-tolerant consensus mechanism based on Raft IMRBFT is proposed.Firstly,the Maglev consensus hash algorithm is used to evenly group the nodes of the blockchain network,and the consensus process is divided into two parts:out-of-group consensus and intra-group consensus.The leader node is first selected in the group,and the node is divided into three levels through the credit mechanism:trusted node,ordinary node and untrusted node.Together with the voting mechanism,it reduces the probability of malicious nodes becoming leader nodes,and forms a committee with other group leader nodes,and the committee selects multiple primary nodes with the highest credit value through the external credit value mechanism to conduct PBFT consensus outside the group.On the basis of Raft consensus,intra-group consensus introduces supervision nodes and relay nodes to further improve security and consensus efficiency and reduce the evil behavior of malicious nodes.Finally,the experimental results show that the communication overhead of IMRBFT increases linearly,the traffic volume is 41.6%of PBFT,the throughput is 4.2 times that of PBFT,and the consensus delay is reduced by 76.4%.With the increase of nodes,the optimization is more obvious,which fully meets the requirements of small communication complexity,high throughput,short consensus delay,and high security and consensus efficiency of large-scale blockchain networks.

blockchainconsensus mechanismnode groupingcredit mechanismsByzantine toleranceRaft algorithm

李莉、李昊泽、李涛

展开 >

东北林业大学计算机与控制工程学院,黑龙江哈尔滨 150040

区块链 共识机制 节点分组 信用机制 拜占庭容错 Raft算法

黑龙江省重点研发计划

2022ZX01A30

2024

广西师范大学学报(自然科学版)
广西师范大学

广西师范大学学报(自然科学版)

CSTPCD北大核心
影响因子:0.448
ISSN:1001-6600
年,卷(期):2024.42(3)
  • 28