Automatic Classification of Network Traffic Data based on Deep Learning in ONOS Platform

Jungmin Kwon, Jungjin Lee, Miseon Yu, Hyunggon Park

Abstract: Machine learning has been deployed in networks for automatically analyzing network data, proactively monitoring network dynamics, and predicting network resource availability. This becomes one of key technologies for efficient and autonomous network management in particular for software defined networks (SDN) environments. Especially, deep learning has brought recent breakthrough in machine learning algorithm as it can extract features based on artificial neural networks from data. In this paper, we study the deployment of deep neural network (DNN) for network traffic data classification, where DNN is deployed to automatically classify real network traffic data collected from ONOS (Open Network Operating System) platform. From the experiment results with simple network topologies, we conclude that DNN can be a potential approach to effective network packet classification. Moreover, it is confirmed that a deployment of DNN for a real network traffic data classification should consider not only the data packets that are intended to be delivered but also data packets required to maintain networks, as the classification performance of DNN significantly depends on the network traffic data.

External IDs:dblp:conf/ictc/KwonLYP20