Computer Engineering and Applications ›› 2021, Vol. 57 ›› Issue (18): 142-148.DOI: 10.3778/j.issn.1002-8331.2006-0046

Previous Articles     Next Articles

Efficient Byzantine Fault Tolerant Algorithm with Supervision Mechanism

WANG Rihong, XING Congying, XU Quanqing, YUAN Shanshan   

  1. 1.School of Information & Control Engineering, Qingdao University of Technology, Qingdao, Shandong 266520, China
    2.Ant Financial Services Group, Hangzhou 310012, China
  • Online:2021-09-15 Published:2021-09-13

具有监督机制的高效拜占庭容错算法

王日宏,邢聪颖,徐泉清,袁杉杉   

  1. 1.青岛理工大学 信息与控制工程学院,山东 青岛 266520
    2.蚂蚁金服,杭州 310012

Abstract:

As the core content of blockchain technology, the consensus mechanism is different in different application fields. In view of the alliance chain application scenario, the widely used Practical Byzantine Fault Tolerancealgorithm (PBFT) still has efficiency and security issues. Therefore, the PBFT algorithm is studied from the perspective of network models, consensus essence and security attacks, and an Efficient Supervised Byzantine Fault Tolerant algorithm(ES-BFT)is proposed. For the efficiency problem, the ES-BFT algorithm randomly divides the nodes into multiple node clusters, sets the reputation value, and selects the consensus node and the supervision node from the node cluster through the reputation value to improve the efficiency and reliability of the consensus node as much as possible. The supervision node is introduced to monitor the consensus node, avoiding the system uncoordinated scenario that the consensus node may encounter before the Global Stabilization Time(GST) begins, and further ensuring the security of the algorithm. Experiments show that the ES-BFT algorithm has improved efficiency and security compared to the PBFT algorithm, and is immune to system incoordination caused by attacks before the GST.

Key words: Practical Byzantine Fault Tolerance(PBFT), Efficiency Supervised Byzantine Fault Tolerant(ES-BFT) algorithm, node clusters, supervisory nodes, Global Stabilization Time(GST)

摘要:

共识机制作为区块链技术的核心内容,在不同应用领域各有差异。针对联盟链应用场景,应用广泛的实用拜占庭容错(PBFT)算法仍然存在效率及安全性问题,因此从网络模型、共识本质及安全攻击等角度对PBFT算法进行研究,提出了一种高效监督拜占庭容错算法(Efficient Supervised Byzantine Fault Tolerance,ES-BFT)。针对效率问题,ES-BFT算法将节点随机划分为多个节点簇,设置信誉值,通过信誉值从节点簇中选举共识节点、监督节点,尽可能提升共识节点的高效性及可靠性;监督节点对共识节点进行监控,避免了在Global Stabilization Time(GST)开始之前共识节点可能遭遇的系统不协调问题,进一步保证算法的安全性;通过实验表明ES-BFT算法在效率及安全性上较PBFT算法有所提升,并且免疫在GST之前的攻击所导致的系统不协调问题。

关键词: 实用拜占庭容错, 高效监督拜占庭容错(ES-BFT)算法, 节点簇, 监督节点, GST