Original title: Network Supervision via Protocol Identification in the Network
Authors: Holasova, E. ; Kuchar, K. ; Fujdiak, R.
Document type: Papers
Language: eng
Publisher: Vysoké učení technické v Brně, Fakulta elektrotechniky a komunikačních technologií
Abstract: This paper is focused on a comparison of ML (Machine Learning) and DNN (Deep Neural Network) techniques in protocol recognition to support network supervision for further proper handling, e.g., detection of a security incident. The DNN approach uses 11 layers and the ML approach is consisting of 28 mutually different predictive models. Both techniques were performed/compared on a freely accessible dataset containing browsing pcap files for further comparison, e.g., with other approaches. The predictive multiclass models were trained (fitted) to be capable of detecting five network protocols. Both approaches were compared by the achieved accuracy (based on testing and validating data), learning time, and predicting the time point of view. Using the ML approach, we were able to recognize the protocol with an accuracy of 1 and using DNN with an accuracy of 0.97.
Keywords: IT protocols; machine learning; neural networks; protocol recognition
Host item entry: Proceedings I of the 28st Conference STUDENT EEICT 2022: General papers, ISBN 978-80-214-6029-4

Institution: Brno University of Technology (web)
Document availability information: Fulltext is available in the Brno University of Technology Digital Library.
Original record: http://hdl.handle.net/11012/209281

Permalink: http://www.nusl.cz/ntk/nusl-524708


The record appears in these collections:
Universities and colleges > Public universities > Brno University of Technology
Conference materials > Papers