PINet: Towards Effective and Efficient Industrial Control Protocol Identification
Jianxin Jin, Wanli Sha, Chaoyang Zhao, Zhan Wang, Yaoyi Zhong, Junliang Zhou · 2023
Accurate identification of industrial control protocols plays a pivotal role in ensuring the secure and reliable operation of Industrial Internet of Things (IIoT) systems. In this paper, we propose, PINet, a novel model that employs a combination of a Gated Recurrent Unit (GRU) network and a multi-head attention mechanism for industrial control protocol identification. PINet effectively captures the underlying patterns within the input sequence by utilizing the hidden state representations obtained from the GRU network. Furthermore, the multi-head attention mechanism facilitates the extraction of crucial information from these hidden states. Ultimately, a classifier linear layer is employed to map the extracted feature vectors onto the probability distribution of protocol categories. Experimental evaluations demonstrate that PINet achieves outstanding performance and accuracy in classifying industrial control protocols, showcasing its effectiveness and reliability in practical applications.