xbeark/traffic_class

PacRep

Source code for KDD'22 paper: "Packet Representation Learning for Traffic Classification".

Requirements

• python: 3.8
• pytorch: 1.8.1
• numpy: 1.19.5
• scikit-learn: 0.24.2
• tensorboard: 2.6.0
• protobuf <= 3.20.0

Usage

Train the model

1. Download preprocessed data from here, and unzip it to ./data/. You can also use your own data with the same format, and change the data path by --data_dir. An example of preprocessing can be found in ./data/pre_exp and preprocess_exp.py.
2. Run the code

python3 run_train.py

3. Results can be found in ./log/sample.log

Reproduce our results

1. Download the trained model. Save the .bin in ./saved_model/, and save the .pth in ./saved_model/sample/
2. Run the code

python3 run_train.py --breakpoint 650

3. Results can be found in ./log/sample.log

Others

Please cite our paper if you use this code in your own work:

@inproceedings{MengWMLLZ22,
  title     = {Packet Representation Learning for Traffic Classification},
  author    = {Xuying Meng and Yequan Wang and Runxin Ma and Haitong Luo and Xiang Li and Yujun Zhang},
  booktitle = {The 28th ACM SIGKDD Conference on Knowledge Discovery and Data Mining},
  pages     = {3546--3554},
  year      = {2022}
}