摘要
共识算法作为区块链的核心技术,决定了区块链系统的性能.其中,PBFT是最具有代表性的一种共识算法,但它存在以下缺点:通信代价大、共识时延长.由此诞生了许多通过各种方法减小共识节点规模来提升PBFT效率的一类算法,但是它们都不是基于距离因素的,并且具有与PBFT类似的缺点.基于此,本文提出了一种基于距离的面向区块链的共识算法.首先通过Grouping算法对节点进行分组,将距离较近的节点分成一组进行共识,从而在缩短共识节点之间距离的基础上减少共识时延.同时结合speculation技术,降低节点间通信的时间复杂度,从而减少了系统的通信代价.最后理论分析了影响共识时延和通信代价的两个指标:节点间的平均延迟、通信次数,证明了DS-PBFT在共识时延以及通信代价这两个方面均要优于PBFT和现有的这一类算法;通过实验验证了DS-PBFT相比PBFT和现有的这一类算法,降低了共识时延.
As the core technology of blockchain,consensus algorithm determines many important characteristics of blockchain system.Among them,PBFT is the most representative consensus algorithm,but it has the following disadvantages:high communication cost and long consensus time.Therefore,many algorithms are born to improve the efficiency of PBFT by reducing the size of consensus nodes by various methods,but they are not based on the distance factor,and have similar disadvantages as PBFT.Based on this,this paper proposes DS-PBFT:a consensus algorithm for blockchain.Firstly,the nodes are grouped by Grouping algorithm,and the nearest nodes are divided into a group for consensus,so as to reduce the consensus delay on the basis of shortening the distance between consensus nodes.At the same time,combining with the speculation technology,the time complexity of communication between nodes is reduced,thus reducing the communication cost of the system.At last,two factors affecting consensus delay and communication cost are theoretically analyzed:average delay between nodes and communication times.It is proved that DS-PBFT is better than PBFT and existing algorithms in consensus delay and communication cost.Experimental results show that DS-PBFT reduces consensus delay compared with PBFT and existing algorithms.
作者
朱海
金瑜
ZHU Hai;JIN Yu(Department of Computer Science and Technology,Wuhan University of Science and Technology,Wuhan 430065,China;Hubei Province Key Laboratory of Intelligent Information Processing and Real-time Industrial,Wuhan 430065,China)
出处
《小型微型计算机系统》
CSCD
北大核心
2022年第3期506-513,共8页
Journal of Chinese Computer Systems
基金
国家自然科学基金项目(61602351,61502359)资助。