Improved PBFT Algorithm Based on Reputation and Voting Mechanism

Abstract

Abstract Aiming at the problems of large communication cost of the consistency process, low efficiency of view switching and the random selection of the master node in the Practical Byzantine Fault Tolerance (PBFT), an Improved Practical Byzantine Fault Tolerance (IPBFT) is proposed. Firstly, This paper proposes a credit model and voting mechanism, that the master node is generated by voting with nodes reputation values, which ensures the reliability of the master node and reduces the possibility that the abnormal node acts as master node. Secondly, the consistency process is optimized to decrease the frequency and time of communication between nodes. Finally, the view switching protocol is optimized, that classification process for abnormal node and malicious node further reduces communication time. The experimental results show that compared with PBFT, IPBFT increases the average throughput, reduces the delay, and improves the operating efficiency and security of the system.