TEST: an End-to-End Network Traffic Examination and Identification Framework Based on Spatio-Temporal Features Extraction.

摘要: With more encrypted network traffic gets involved in the Internet, how to effectively identify has become a top priority field. Accurate identification of is footstone basic services, say QoE, bandwidth allocation, and IDS. Previous methods either cannot deal with traffics or require experts select tons features attain relatively decent accuracy.In this paper, we present Deep Learning based end-to-end framework, termed TEST, avoid aforementioned problems. CNN LSTM are combined implemented help machine automatically extract from both special time-related raw traffic. The presented framework two layers structure, which made it possible remarkable accuracy on classification intrusion detection tasks. experimental results demonstrate that our model can outperform previous state-of-the-art 99.98%.