RR-Raft: A Raft Consensus Algorithm Based on Reputation Regression Model and RSA Signature

To address the issue that existing Raft consensus algorithms in blockchain cannot achieve low latency, high security, and high efficiency simultaneously, a Raft consensus algorithm based on reputation regression and RSA signatures, named RR-Raft, is designed. The algorithm first introduces a reputation measurement function and a reputation consumption function into the Logistic regression algorithm, constructing a reputation model based on R-Logistic regression. This model solves the centralized reputation problem in traditional regression algorithms, thereby enhancing the enthusiasm of nodes for consensus. Secondly, to address the issue of vote splitting during the leader election phase of the consensus mechanism, a quick leader node election strategy based on the reputation model is proposed to improve consensus efficiency. Lastly, to prevent the Leader node from tampering with the logs as a malicious actor, a fast RSA aggregate signature based on Barrett reduction is proposed to enhance the security of consensus.