Abstract:Traffic classification occupies a significant role in cybersecurity and network management. The widespread of encryption transmission protocols such as SSL/TLS has led to the dominance of deep learning based approaches. In cybersecurity, strong adversaries often complicate their strategies by constantly developing emerging attacks. Meanwhile, security practitioners desire to grasp the reasons for inference results. However, existing deep learning approaches lack efficient adaptation for incremental traffic types and often have less interpretability. In this paper, we propose I <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$^{2}$</tex-math></inline-formula> RNN, an Incremental and Interpretable Recurrent Neural Network for encrypted traffic classification. The I <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$^{2}$</tex-math></inline-formula> RNN proposes a novel propagation process to extract the sequence fingerprints from sessions with local robustness. Meanwhile, this proposal provides interpretability including time-series feature attribution and inter-class similarity portrait. Moreover, we design I <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$^{2}$</tex-math></inline-formula> RNN in an incremental manner to adapt to emerging traffic types. The I <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$^{2}$</tex-math></inline-formula> RNN only needs to train an additional set of parameters for the newly added traffic type rather than retraining the whole model with the entire dataset. Extensive experimental results show that our I <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$^{2}$</tex-math></inline-formula> RNN can achieve remarkable performance in traffic classification, incremental learning, and model interpretability. Compared with other local interpretability methods, our I <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$^{2}$</tex-math></inline-formula> RNN exhibits excellent stability, robustness, and effectiveness in the interpretation of network traffic data.

MTT: an efficient model for encrypted network traffic classification using multi-task transformer

Network Traffic Classification Via Non-Convex Multi-Task Feature Learning

MATEC: A lightweight neural network for online encrypted traffic classification

I <inline-formula><tex-math notation="LaTeX">$^{2}$</tex-math></inline-formula> RNN: An Incremental and Interpretable Recurrent Neural Network for Encrypted Traffic Classification

MTC-NET: A Multi-Channel Independent Anomaly Detection Method for Network Traffic

NetMamba: Efficient Network Traffic Classification via Pre-training Unidirectional Mamba

A Multi-Modal Approach For Context-Aware Network Traffic Classification

Multitask Learning for Network Traffic Classification

Novel Dynamic Multiple Classification System for Network Traffic

L2-BiTCN-CNN: Spatio-temporal features fusion-based multi-classification model for various internet applications identification

Towards a Transformer-Based Pre-trained Model for IoT Traffic Classification

Identification of Encrypted Traffic Through Attention Mechanism Based Long Short Term Memory

Network Traffic Classification Model Based on Spatio-Temporal Feature Extraction

Multi-Task Scenario Encrypted Traffic Classification and Parameter Analysis

Multa: Enabling Cost-efficient Multi-task Network Traffic Analysis

An encrypted traffic identification method based on multi-scale feature fusion

TCGNN: Packet-grained Network Traffic Classification Via Graph Neural Networks

Lightweight Transformer Model for Mobile Application Classification

CM-UTC: A Cost-sensitive Matrix based Method for Unknown Encrypted Traffic Classification

Encrypted Network Traffic Classification Using Deep And Parallel Network-In-Network Models

End-to-end encrypted traffic classification with one-dimensional convolution neural networks