【摘要】 实用拜占庭容错共识算法（PBFT）作为联盟链中最常见的共识算法，可以在恶意节点少于三分之一的情况下，保证系统的正确性.然而，PBFT算法在建立信任与共识过程中存在高时延、低吞吐量、主节点选取安全性、恶意节点未处理等问题.为了解决这些问题，引入奖励积分机制来对共识节点进行分组并设定候选节点集，提出了一种高效快速的拜占庭容错算法（DIG-PBFT）.在共识过程中，DIG-PBFT通过动态地调整实际参与共识的节点，增加了安全性更高节点的参与度.仿真实验结果表明：与PBFT算法及其同类工作相比，DIG-PBFT算法的吞吐量更大、时延更低，且安全性更高.更多还原

【Abstract】 Practical Byzantine Fault Tolerance Consensus Algorithm(PBFT), as the most common consensus algorithm in alliance chains, can ensure the correctness of the system when there are less than one third of the malicious nodes.However, the PBFT algorithm has problems such as high latency, low throughput, security of master node selection, and unprocessed malicious nodes in the process of establishing trust and consensus. In order to solve these problems, a reward point mechanism is introduced to group consensus nodes and sets a candidate node set, and proposes an efficient and fast Byzantine fault tolerant algorithm(DIG-PBFT). During the consensus process, DIG-PBFT increases the participation of nodes with higher security by dynamically adjusting the nodes that actually participate in the consensus. The simulation experiment results show that compared with the PBFT algorithm and its similar work, the DIG-PBFT algorithm has greater throughput, lower latency, and higher security.更多还原