TEST: an End-to-End Network Traffic Classification System With Spatio-Temporal Features Extraction
Yi Zeng, Zihao Qi, Wencheng Chen, Yanzhe Huang · 2019
With more encrypted network traffic gets involved in the Internet, how to effectively classify network traffic has become a more and more important task in the field. Accurate classification of the network traffic is the footstone of basic network services, say QoE, bandwidth allocation, and Intrusion Detection System (IDS). Previous classification methods have many shortcomings since they cannot deal with the encrypted network traffic but require human experts to select tons of features to attain a relatively decent accuracy. Therefore, in this paper, we present a Deep Learning based end-to-end network traffic classification framework, termed TEST, to avoid the aforementioned problems. CNN and LSTM are combined and implemented to help the network intrusion detection systems automatically extract features from both special and time-related features of the raw traffic. The structure of our framework has two layers, which made it possible to attain a remarkable accuracy on encrypted traffic classification tasks. Compared to the LeNet model with the accuracy of 80.27% and the LSTM model with the accuracy of 81.96%, terimental results demonstrate that our model can outperform previous methods with a state-of-the-art accuracy of 99.98%.