PB-Raft: A Byzantine fault tolerance consensus algorithm based on weighted PageRank and BLS threshold signature

Abstract Addressing the challenge of balancing consensus efficiency and Byzantine fault tolerance in prevalent consensus algorithms within consortium blockchains, this paper proposes an improved Raft algorithm——PB-Raft, based on the weighted PageRank algorithm and BLS threshold signature. First, it improves the leader election process through the weighted PageRank algorithm, constructing a weighted directed graph among nodes, computing the PageRank values of each node, and reducing the election timeout for node with the highest PageRank values to improve leader election efficiency and thereby enhance algorithm's consensus efficiency. Then, it improves the log replication process through BLS threshold signature, transforming the log replication process into a threshold signature verification process, achieving Byzantine fault tolerance during the log replication process, and hence improving the algorithm's security. Finally, the PB-Raft algorithm was simulated in Go language on Hyperledger Fabric, and through multiple tests using Caliper, it was shown that compared to the Raft algorithm, PB-Raft's leader election efficiency increases by approximately 30\%, and it has Byzantine fault tolerance capability, increases throughput by about 10\%, and reduces consensus delay by about 30\%. These results demonstrate that the PB-Raft algorithm not only possesses Byzantine fault tolerance capability but also enhances consensus efficiency, offering new insights and viable solutions for the design of consortium blockchain consensus algorithms.