Genetic Programming Based Network Traffic-Profiling System

  • M. Ekmanis Riga Technical University

Abstract

A new approach is proposed for unattended grouping of traffic sources. This is GA based search technique, which use finite state machine as a genetic representation of solution domain. Models approximate multi host communication in relation to host under research. They describe traffic source behavior handling event sequence with internal loops and provide similar traffic source grouping. The proof of concept implementation shows potential capability to detect network anomalies and sources of unwanted traffic. Ill. 7, bibl. 8 (in English; summaries in English, Russian and Lithuanian).

Published
2015-03-10
Section
T 180 TELECOMMUNICATION ENGINEERING