Grouping improvement of PBFT algorithm based on signature

Because the traditional consensus algorithm has the problems of low fault tolerance, low throughput and high delay, it cannot support large-scale distributed systems. In addition, the performance of some algorithms can not meet the requirements of modern Internet applications. In order to solve these limitations, researchers have proposed several improvement schemes. Firstly, the network is divided into groups based on delay and reputation mechanism, and a blockchain network structure of master-slave multi-chain is constructed. When there are many participating nodes, we use HotStuff algorithm, which has the characteristics of high scalability and high performance. By using different consensus algorithms in different scenarios, we can not only ensure the strong fault tolerance of the system, but also improve the throughput and efficiency of the system by dividing the fragments based on delay and improving the consensus algorithm. We have carried out experiments on the blockchain experimental platform, and verified that the algorithm has good performance and fault tolerance. By comparing with other frameworks, we find that the algorithm has better performance in different scenarios. This research has important reference significance for building efficient and reliable distributed systems.