BFTDiagnosis: An automated security testing framework with malicious behavior injection for BFT protocols

Jitao Wang,Bo Zhang,Kai Wang,Yuzhou Wang,Weili Han

DOI: https://doi.org/10.1016/j.comnet.2024.110404

IF: 5.493

2024-05-10

Computer Networks

Abstract:In peer-to-peer computing networks, Byzantine Fault Tolerance (BFT) protocols are a popular solution to ensure the consistency security in the presence of some malicious nodes, thus are widely employed in blockchain systems. However, BFT protocols still face various security threats in practice. Currently, testing the security of BFT protocols often requires manual operations. I.e., it lacks comprehensive automation testing techniques. This makes it difficult to cover various scenarios of malicious node behaviors. Therefore, it is an urgent requirement to design an automated testing framework that can comprehensively and efficiently evaluate the security of BFT protocols. This paper proposes BFTDiagnosis , an automated security testing framework with malicious behavior injection for BFT protocols. The framework can automatically configure and initiate protocol testing tasks, and construct consensus nodes to execute different BFT protocols. During testing, consensus nodes can inject malicious behaviors based on pattern matching strategies and malicious behavior triggering mechanisms to simulate various malicious scenarios. Through an Analyzer, BFTDiagnosis can collect protocol runtime data from consensus nodes and calculate four security quantification indicators to evaluate protocol performance. We conducted a load consumption evaluation of BFTDiagnosis . The results indicate its acceptable performance. Additionally, we tested PBFT, Basic-HotStuff, and Chained-HotStuff protocols using the BFTDiagnosis framework, validating the effectiveness of the framework and the rationality of security quantification indicators. Finally, we compared it with nine related technologies and found that BFTDiagnosis is good at testing scenario comprehensiveness and fault localization capability, and can intuitively evaluate the security of BFT protocols through four quantification indicators. The above results show that BFTDiagnosis is an effective and comprehensive security testing framework for BFT protocols.

computer science, information systems,telecommunications,engineering, electrical & electronic, hardware & architecture