【摘要】 【目的】为解决实用拜占庭容错算法(practical Byzantine fault-tolerant algorithm, PBFT)通信复杂度高、共识时延高等不足，提出节点间通信时延分组的改进实用拜占庭容错算法(grouping PBFT,GPBFT)。【方法】首先将区块链系统节点进行分组，依据最少网络通信次数确定分组数；然后计算各组节点间平均通信时延进行组内节点筛选，确定组内节点数；最后以共识成功率、失败率和节点历史行为评估参数为变量计算节点信誉值，监督节点共识行为，减少异常节点的参与。【结果】通过基于Hyperledger Fabric平台的区块链系统进行仿真试验，结果表明：与PBFT相比，GPBFT平均时延降低57.86%、平均吞吐量提高55.04%,通信复杂度数量级由平方级降低为对数级。【结论】GPBFT可满足多节点场景下区块链复杂通信的高时效性需求，解决了行业区块链系统大规模节点的需求问题。更多还原

【Abstract】 [Objective] In response to the shortcomings of high communication complexity and high consensus delay that plague the practical Byzantine fault-tolerant algorithm(PBFT), an improved practical Byzantine fault-tolerant algorithm(grouping PBFT, GPBFT) was proposed for grouping inter-node communication delay. [Method] First, blockchain system nodes were grouped, and the number of groups was determined according to the minimum number of network communication; then, the average communication delay between each group of nodes was calculated to screen the nodes in the group and determine the number of nodes in the group; finally, the consensus success rate, the failure rate and the historical behavior evaluation parameters of nodes were used as variables to calculate the reputation values of nodes, supervise the consensus behavior of nodes, and reduce the participation of abnormal nodes. [Result] Through the simulation test of the blockchain system based on the Hyperledger Fabric platform, the results show that compared with PBFT, the average delay of GPBFT decreases by 57.86%, the average throughput increases by 55.04%, and the order of magnitude in communication complexity decreases from the square to logarithmic. [Conclusion] GPBFT can meet the high timeliness requirements of complex blockchain communication in multi-node scenarios, and fulfill the needs of large-scale nodes of the industry blockchain system.更多还原