基于哈希值分组和信任主节点选取的共识机制

Consensus Mechanism Based on Hash Grouping and Trust Primary Node Selection

PBFT(Practical Byzantine Fault Tolerance)算法是当前最流行的共识机制之一,但其采用广播的通信模式导致该算法具有高通信复杂度;将节点分层进行共识的改进方式虽降低了通信规模,但改进后的算法在安全性和效率等方面仍存在不足。基于此,提出一种HBFT改进算法。首先,根据节点MAC地址的哈希值进行分组,增加节点分组的随机性并使拜占庭节点的分布更均匀;其次,优化算法的共识流程、通信内容和视图更换协议,进一步降低通信规模并提高主节点更换和故障处理的速度;最后,引入信誉机制并据此选取主节点,提升主节点的可靠性和算法的安全性。从理论、实验和安全的角度进行分析验证,结果表明HBFT算法的共识效率较PBFT算法和基于分层的改进算法分别提高96.1%~98.6%,51.3%~89.7%,且HBFT算法具有更高的安全性。

Practical Byzantine Fault Tolerance algorithm(PBFT)is one of the most popular consensus mechanisms.However,the communication mode of its broadcast leads to high communication complexity.Although the improved method of hierarchical node consensus reduces the communication scale of the algorithm,there are still shortcomings in security and efficiency of the algorithm.Based on this,an improved HBFT algorithm is proposed.Firstly,the nodes are grouped according to the hash value of MAC address,which increases the randomness of node grouping and makes the distribution of Byzantine nodes more uniform.Secondly,the consensus process,communication content and view replacement protocol of the algorithm are optimized to further reduce the communication scale and improve the speed of master node replacement and fault handling.Finally,the credit mechanism is introduced and the master node is selected based on it.It improves the reliability of the master node and the security of the algorithm.From the perspective of theory,experiment and safety,it is showed that the consensus efficiency of HBFT algorithm is improved by 96.1%~98.6%and 51.3%~89.7%compared with PBFT algorithm and layer-based improved algorithm.In addition,HBFT algorithm has higher security.