WPBFT:An improved consensus algorithm based on the HotStuff algorithm
Mingan Gao, Gehao Lu, Zhiyuan Wang, Yulian Gao · 2023
This paper first analyzes the consensus algorithm that currently exists in blockchains, and then mainly discusses the PBFT consensus algorithm widely used in consortium chains and the improved HotStuff algorithm based on it. However, they also have some disadvantages, such as low throughput, especially when nodes continue to increase, throughput will decline rapidly. Although HotStuff has improved the masternode election mechanism, the signature encryption algorithm it uses is insecure and inefficient. Finally, based on these problems, we propose a consensus algorithm WPBFT based on HotStuff’s improvement. Our consensus algorithm is improved in the following ways. First, the new consensus mechanism will not need to elect a masternode, but will use random selection of masternodes, so we will change the time complexity of HotStuff from O(n) to O(1), greatly reducing network communication. Secondly, this paper adopts the Ed25519 encryption algorithm, and improves its foundation, introducing a large number compression algorithm, which effectively improves the security and encryption efficiency when the message type is relatively large. Finally, it adopts a chronologically-based chain structure at the nodes. Through experiments, we show that the proposed algorithm has been significantly improved compared with PBFT and HotStuff algorithms in terms of data throughput and communication delay.