基于Raft投票机制的PBFT改进共识算法RBFT

Improved Consensus Algorithm RBFT for PBFT Based on Raft Voting Mechanism

PBFT算法存在着时延长、效率不高、动态性不高的问题,论文提出一种基于Raft投票机制共识算法RBFT(Raft Byzantine Fault Tolerant)。由于PBFT算法中主节点选取过于随意,易恶意节点易重新当选,该算法将节点进行分层,分为共识域和备份域,主节点选取改为Raft中Leader的投票选举,在共识域R1共识的过程中,同时进行备选域R2备选主节点(Leader)的选举,实现了备份域节点的动态增减,同时优化视图切换协议。减少了通信消耗并进一步降低时延。经实验后得出,改进后算法与原有算法相比,时延有了进一步的降低,吞吐量得到更大的提升。

PBFT algorithm has the problems of time extension,inefficiency and dynamism,this paper proposes a consensus algorithm RBFT(Raft Byzantine Fault Tolerant)based on Raft voting mechanism.Since the selection of master nodes in PBFT algo⁃rithm is too arbitrary and easy to re-elect malicious nodes,this algorithm layers the nodes into consensus domain and backup do⁃main,the selection of master nodes is changed to the voting election of Leader in Raft,and the election of alternative master nodes(Leader)in alternative domain R2 is conducted simultaneously in the process of consensus in consensus domain R1,which realizes the dynamic increase and decrease of nodes in backup domain,and at the same time,the view switching protocol in optimized.The communication consumption is reduced and the latency is further reduced.After experiments,it is concluded that the improved algo⁃rithm has further reduced the latency and improved the throughput compared with the original algorithm.