A Practical Byzantine Fault-tolerant Consensus Algorithm Based On Dual Administrator Short Group Signatures

Blockchain(Blockchain)is a new application model of computer technology such as distributed storage,peer-to-peer transmission,consensus mechanism,and cryptographic algorithm,whose core features are decentralization and tamper-evident.One of the processes to solve the consistency problem in blockchain is called consensus.Consensus reaching in distributed systems needs to rely on reliable consensus algorithms,which usually solve the problem of which node initiates a proposal in a distributed system and how other nodes agree on this proposal.At present,consensus algorithms can be divided into three types according to the type of blockchain: public chain consensus,private chain consensus and coalition chain consensus,among which coalition chain is more suitable for various application scenarios,and Practical Byzantine Fault Tolerance(PBFT),as one of the core elements of coalition chain,has also become a hot research topic.The Practical Byzantine Fault Tolerance(PBFT)algorithm optimizes the operational complexity of the algorithm to polynomial level through a three-stage protocol,so that the BFT-type algorithm can also be applied in distributed systems,but it still has problems such as arbitrary selection of master nodes,high communication overhead,low security,poor scalability,and difficulty in tracking and other problems.In this paper,a Practical Byzantine Fault Tolerance(GPBFT)consensus algorithm based on short group signatures of dual administrators is proposed to address these problems.The main research elements of the algorithm are as follows:Firstly,in response to the problem of node selection in the PBFT algorithm,the paper selects a decision tree model to select various nodes in the algorithm.The first step is to process a dataset about node features,and then construct a decision tree node classification model based on this dataset.Based on the selection results of the model,the main node and the group administrator with short group signatures are selected.The main node and group member administrators are the same node,and the group tracking administrator is selected separately.If the first entered node in the algorithm meets the various features selected by the model,it becomes the main node of the current consensus cycle,and the main node will be re selected at the beginning of each round of consensus process.Secondly,a PBFT algorithm based on double administrator short group signature,namely the GPBFT algorithm,was proposed.In the Group Sign stage of the GPBFT algorithm,group member administrators organize nodes into a group and perform short group signatures.The messages sent by any group member node can represent the entire group,thereby reducing a large amount of communication overhead between nodes.In addition,the GPBFT algorithm also adds a Trace stage,where group tracking administrators can track the true identity of malicious nodes based on requirements and client verification results,determine the information of malicious nodes,and revoke them.After experimental verification,the GPBFT algorithm reduces the network communication overhead and consensus latency compared with the traditional PBFT algorithm in the case of multiple nodes,which greatly improves the security and stability of the system.The algorithm can effectively manage member nodes and enable the tracking of identified malicious nodes while maintaining anonymity in terms of node tracking.