Journal
COMPUTER NETWORKS
Volume 200, Issue -, Pages -Publisher
ELSEVIER
DOI: 10.1016/j.comnet.2021.108535
Keywords
Encrypted traffic classification; Generative Adversarial Network; Semi-supervised learning; Traffic identification; SDN gateway
Ask authors/readers for more resources
With the rapid development of communication network technology, network traffic data types and quantities are increasing, making network traffic classification more important. SDN Edge Gateway, as the closest network element to users, can greatly enhance user experience through traffic classification capabilities. Utilizing GAN-based semi-supervised learning methods can improve traffic classification performance without the need for large labeled datasets.
With the rapid development of communication network technology, the types and quantity of network traffic data are accordingly increasing. Network traffic classification has become a non-trivial research task in the area of network management and security, which not only helps to improve the fine-grained network resource allocation but also enables policy-driven network management. As the closest network element to the subscribers, SDN Edge Gateway can tremendously enhance the user experience with the capability of traffic classification. Deep Learning (DL) performs automatic feature extraction without human intervention, which undoubtedly makes it a highly desirable approach for traffic classification, especially encrypted traffic. However, capturing large labeled datasets is cumbersome and time-consuming manual labor. Semi-Supervised learning is an appropriate technique to overcome this problem. With that in mind, we proposed a Generative Adversarial Network (GAN)-based Semi-Supervised Learning Encrypted Traffic Classification method called ByteSGAN embedded in SDN Edge Gateway to achieve the goal of traffic classification in a fine-grained manner to further improve network resource utilization. ByteSGAN can only use a small number of labeled traffic samples and a large number of samples to achieve a good performance of traffic classification by modifying the structure and loss function of the regular GAN discriminator network in a semi-supervised learning way. Based on the public datasets 'ISCX2012 VPN-non VPN' and 'Crossmarket', two experimental results show that the ByteSGAN can efficiently improve the performance of traffic classifiers and outperform the other supervised learning methods like CNN.
Authors
I am an author on this paper
Click your name to claim this paper and add it to your profile.
Reviews
Recommended
No Data Available