C-GRU:A Parallel Neural Network for Malicious Traffic Classification
Bo Wang, Yang Su, Junke Nie · 2020
The network is a double-edged sword. While people enjoy the convenience brought by the development of network communication technology, various malicious behaviors hidden in the network invade users gradually. The intrusion detection system (IDS) is designed to detect the intrusion of the outside world to the system and make corresponding emergency response. Flow detection is often an important part of the IDS. However, traditional IDS are based on experts manually designing intrusion features, and feature library maintenance costs are high and new-type intrusion behaviors cannot be detected. To this end, based on the spatial-temporal characteristics of network traffic, we designed a parallel structure convolutional neural network that combines convolutional neural network (CNN) and gated recurrent unit (GRU) to classify malicious traffic. We evaluated the neural network model we designed on the public ISCX2012 dataset. The experimental results show that our model can effectively classify the type of traffic and avoid the occurrence of false alarms.