ENIPFuzz: A SeqGAN-based EtherNet/IP Protocol Fuzzing Test Framework
Honggang Wu, Li Gong, Ao Liu, Yi Zhang, Jianwei Yang · 2022 IEEE 5th International Conference on Electronics Technology (ICET) · 2022
How to effectively dig out the potential vulnerabilities of communication protocols is essential for improving the security of the industrial control systems (ICS). However, the traditional fuzzing methods has some limitations for the digging of new vulnerabilities. In addition, there are few EtherNet/IP protocol vulnerabilities in the public vulnerability database. In this paper, we introduce an EtherNet/IP protocol fuzzing test framework based on the Sequence Generation Adversarial Network (SeqGAN), which can automatically learn protocol grammar and generate test cases. We contributions are twofold. First, we propose a SeqGAN-based EtherNet/IP protocol fuzzing test framework (ENIPFuzz), and this framework does not rely on existing vulnerability database. Then, in order to increase the diversity of generated test cases, we propose three clustering strategies to construct the training set. Besides, to prove the usability of our framework, we deployed ENIPFuzz to the ICS that using the EtherNet/IP protocol and discovered several vulnerabilities.