Murdoch University Research Repository

Welcome to the Murdoch University Research Repository

The Murdoch University Research Repository is an open access digital collection of research
created by Murdoch University staff, researchers and postgraduate students.

Learn more

Internet traffic classification using an ensemble of deep convolutional neural networks

Shahraki, A., Abbasi, M., Taherkordi, A. and Kaosar, M. (2021) Internet traffic classification using an ensemble of deep convolutional neural networks. In: 4th FlexNets Workshop on Flexible Networks Artificial Intelligence Supported Network Flexibility and Agility (Part of SIGCOMM '21), 23 August 2021, Virtual pp. 38-43.

Free to read:
*No subscription required


Network traffic classification (NTC) has attracted considerable attention in recent years. The importance of traffic classification stems from the fact that data traffic in modern networks is extremely complex and ever-evolving in different aspects, e.g. volume, velocity and variety. The inherent security requirements of Internet-based applications also highlights further the role of traffic classification. Gaining clear insights into the network traffic for performance evaluation and network planning purposes, network behavior analysis, and network management is not a trivial task. Fortunately, NTC is a promising technique to gain valuable insights into the behavior of the network, and consequently improve the network operations. In this paper, we provide a method based on deep ensemble learning to classify the network traffic in communication systems and networks. More specifically, the proposed method combines a set of Convolutional Neural Network (CNN) models into an ensemble of classifiers. The outputs of the models are then combined to generate the final prediction. The results of performance evaluation show that the proposed method provides an average accuracy rate of 98% for the classification of traffic (e.g., FTP-DATA, MAIL, etc.) in the Cambridge Internet traffic dataset.

Item Type: Conference Paper
Murdoch Affiliation(s): IT, Media and Communications
Item Control Page Item Control Page