TU Darmstadt / ULB / TUbiblio

Unsupervised Traffic Flow Classification Using a Neural Autoencoder

Höchst, Jonas and Hollick, Matthias and Freisleben, Bernd and Baumgärtner, Lars :
Unsupervised Traffic Flow Classification Using a Neural Autoencoder.
In: 42nd Annual IEEE Conference on Local Computer Networks (LCN 2017), Singapore, Singapore. Singapore, Singapore
[Conference or Workshop Item] , (2017)

Abstract

The delay and bandwidth requirements of today's mobile applications differ widely depending on the functionality provided and the data transferred, such as uploading user generated multimedia content, playing online games, or streaming live videos. To perform resource management, mobile wireless networks can particularly profit from classifying and predicting mobile application traffic. State-of-the-art traffic classification approaches have various disadvantages: port-based classification methods can be circumvented by choosing non-standard ports, protocol fingerprinting can be confused by the use of encryption, and current supervised learning methods for analyzing the statistical properties of network flows try to detect predefined classes, such as e-mail or FTP traffic, learned during training. In this paper, we present a novel approach to unsupervised traffic flow classification using statistical properties of flows and clustering based on a neural autoencoder. In contrast to previous work, the neural autoencoder is used to automatically cluster traffic flows, for example, into downloads, uploads, or voice calls, independent of the particular network protocols, such as FTP or HTTP(S), used for performing these tasks. A novel time interval based feature vector construction and a semi-automatic cluster labeling method facilitate traffic flow classification independent of known traffic classes or classification constraints. An experimental evaluation of the proposed approach on real data captured from about 25 mobile devices performing daily work over a period of 4 months is presented. The obtained results show that 7 different classes of mobile traffic flows are detected with an average precision of 80% and an average recall of 75%, indicating the feasibility of our approach.

Item Type: Conference or Workshop Item
Erschienen: 2017
Creators: Höchst, Jonas and Hollick, Matthias and Freisleben, Bernd and Baumgärtner, Lars
Title: Unsupervised Traffic Flow Classification Using a Neural Autoencoder
Language: English
Abstract:

The delay and bandwidth requirements of today's mobile applications differ widely depending on the functionality provided and the data transferred, such as uploading user generated multimedia content, playing online games, or streaming live videos. To perform resource management, mobile wireless networks can particularly profit from classifying and predicting mobile application traffic. State-of-the-art traffic classification approaches have various disadvantages: port-based classification methods can be circumvented by choosing non-standard ports, protocol fingerprinting can be confused by the use of encryption, and current supervised learning methods for analyzing the statistical properties of network flows try to detect predefined classes, such as e-mail or FTP traffic, learned during training. In this paper, we present a novel approach to unsupervised traffic flow classification using statistical properties of flows and clustering based on a neural autoencoder. In contrast to previous work, the neural autoencoder is used to automatically cluster traffic flows, for example, into downloads, uploads, or voice calls, independent of the particular network protocols, such as FTP or HTTP(S), used for performing these tasks. A novel time interval based feature vector construction and a semi-automatic cluster labeling method facilitate traffic flow classification independent of known traffic classes or classification constraints. An experimental evaluation of the proposed approach on real data captured from about 25 mobile devices performing daily work over a period of 4 months is presented. The obtained results show that 7 different classes of mobile traffic flows are detected with an average precision of 80% and an average recall of 75%, indicating the feasibility of our approach.

Place of Publication: Singapore, Singapore
Uncontrolled Keywords: Autoencoder; Traffic flow classification; Unsupervised learning
Divisions: Department of Computer Science > Sichere Mobile Netze
DFG-Collaborative Research Centres (incl. Transregio) > Collaborative Research Centres > CRC 1053: MAKI – Multi-Mechanisms Adaptation for the Future Internet
DFG-Collaborative Research Centres (incl. Transregio) > Collaborative Research Centres > CRC 1053: MAKI – Multi-Mechanisms Adaptation for the Future Internet > A: Construction Methodology > Subproject A3: Migration
DFG-Collaborative Research Centres (incl. Transregio) > Collaborative Research Centres
DFG-Collaborative Research Centres (incl. Transregio) > Collaborative Research Centres > CRC 1053: MAKI – Multi-Mechanisms Adaptation for the Future Internet > A: Construction Methodology
Department of Computer Science
DFG-Collaborative Research Centres (incl. Transregio)
Event Title: 42nd Annual IEEE Conference on Local Computer Networks (LCN 2017)
Event Location: Singapore, Singapore
Date Deposited: 19 Oct 2017 12:33
Export:

Optionen (nur für Redakteure)

View Item View Item