结合信誉积分的PBFT改进方案

Improved PBFT Scheme Combined with Credit Score

为解决目前PBFT算法中主节点作恶和一次共识时间较长的问题,提出一种基于双重信誉机制的实用拜占庭容错算法(DB-PBFT)。首先引入节点的动态信誉值和静态信誉值得到节点的总信誉值,以此来标记节点状态,对节点进行划分;其次根据划分后的结果选择主节点并建立共识群组,减少参与共识的节点数目;最后对引入奖惩机制,对活跃度低的节点降低其信誉度,对于作恶节点直接将其踢出共识群组以此来维护网络的安全。实验表明,DB-PBFT算法比起PBFT算法拥有更高的效率、吞吐量和低能耗性。

In order to solve the problem that the main node has committed crimes and a long time of consensus in the current PBFT,a Practical Byzantine Fault Tolerant Algorithm based on dual reputation mechanism(DB-PBFT)was proposed.First,the dynamic reputation value and static reputation value of the node are introduced to get the total reputation value of the node,which is used to mark the node status and divide the node.Secondly,according to the results of the division,select the master node and establish a consensus group to reduce the number of nodes participating in consensus.Finally,the introduction of reward and punishment mechanism,and the malicious node is directly kicked out of the consensus group to maintain the security of the network.Experiments show that DB-PBFT algorithm has higher efficiency,throughput and low energy consumption than PBFT algorithm.