Deep packet: a novel approach for encrypted traffic classification using deep learning

By Mohammad Lotfollahi, Mahdi Jafari Siavoshani, Ramin Shirali Hossein Zade, Mohammdsadegh Saberian. Soft Computing - A Fusion of Foundations, Methodologies and Applications, 2017

1,082 citations recorded by the source corpus.

Network traffic classification has become more important with the rapid growth of Internet and online applications. Numerous studies have been done on this topic which have led to many different approaches. Most of these approaches use predefined features extracted by an expert in order to classify network traffic. In contrast, in this study, we propose a deep learning-based approach which integrates both feature extraction and classification phases into one system. Our proposed scheme, called “Deep Packet,” can handle both traffic characterization in which the network traffic is categorized into major classes (e.g., FTP and P2P) and application identification in which identifying end-user applications (e.g., BitTorrent and Skype) is desired. Contrary to most of the current methods, Deep Packet can identify encrypted traffic and also distinguishes between VPN and non-VPN network traffic. The Deep Packet framework employs two deep neural network structures, namely stacked autoencoder (SAE) and convolution neural network (CNN) in order to classify network traffic. Our experiments show that the best result is achieved when Deep Packet uses CNN as its classification model where it achieves recall of 0.98 in application identification task and 0.94 in traffic categorization task. To the best of our knowledge, Deep Packet outperforms all of the proposed classification methods on UNB ISCX VPN-nonVPN dataset.

Subfields: Artificial Intelligence, Computer Networks and Communications, Signal Processing

View Paper: https://www.semanticscholar.org/paper/36f13179cdfc13017df535fdee582d58067301f3

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

1. Gang Xiong
2. Paul Patras
3. Fei Hu
4. Qian Mao
5. Zhen Li
6. Hamed Haddadi
7. Qi Hao
8. Domenico Ciuonzo

In Computer Networks and Communications

1. Giuseppe Aceto
2. Antonio Pescapè
3. Paul Patras
4. Fei Hu
5. Qian Mao
6. Hamed Haddadi
7. Qi Hao
8. Domenico Ciuonzo

In Signal Processing

1. Gang Xiong
2. Paul Patras
3. Fei Hu
4. Qian Mao
5. Zhen Li
6. Hamed Haddadi
7. Qi Hao
8. Domenico Ciuonzo

Citations in the corpus

• Deep Learning for Encrypted Traffic Classification: An Overview (2018, positive): “In some recent DL-based approaches, entire packets are taken as input [6].”
• FS-Net: A Flow Sequence Network For Encrypted Traffic Classification (2019, methodology): “Lotfollahi et al. (2017) adopted the stacked autoencoder and one-dimensional convolution neural network to extract features from encrypted traffic payloads automatically [25].”
• ET-BERT: A Contextualized Datagram Representation with Pre-training Transformers for Encrypted Traffic Classification (2022, methodology): “In contrast, traditional supervised methods (e.g. BIND, DF, FS-Net) show substantial F1 performance degradation when the sample size is reduced, e.g. Deeppacket’s performance decreases by 40.22% when the sample size is reduced from the full size to 10%.”
• Deep Learning for Intelligent Wireless Networks: A Comprehensive Survey (2018, methodology): “[101] proposed a DL-based traffic classification method, namely, deep packet, which not only distinguishes traffic type (such as streaming, P2P, etc.”
• Mobile Encrypted Traffic Classification Using Deep Learning: Experimental Evaluation, Lessons Learned, and Challenges (2019, methodology): “Differently, in [22] a 1D-CNN consisting of two 1D convolutional layers (200 and 80 filters, respectively, with 1D average-pooling) and seven fully-connected layers (with”
• A Survey of Deep Learning Methods for Cyber Security (2019, methodology): “[162] extended Wang’s work by exploring traffic characterization and application identification.”
• Deep Learning in Mobile and Wireless Networking: A Survey (2018, methodology): “present Deep Packet, which is based on a CNN, for encrypted traffic classification [199].”
• Towards the Deployment of Machine Learning Solutions in Network Traffic Classification: A Systematic Survey (2019, neutral): “A novel approach to identify encrypted applications is given by [170], which presents the traffic classification problem”