In 2014, we launched our open-access repository which offers full text access to conference proceedings from many of our events including the INC and HAISA series. These papers are free to access and distribute (subject to citing the source).
Ninth International Network Conference (INC 2012)
Title: Establishing discernible flow characteristics for accurate, real-time network protocol identification
Author(s): Ryan Goss, Reinhardt Botha
Keywords: Network Traffic Flows, Deep Packet Inspection, Traffic Classification, Statistical Profiling, Flow Discrimination, Artificial Neural Network
Abstract: As the number of business applications, games and other real-time applications adopting peer-to-peer communication approaches increase, so too does the requirement for accurate identification of these protocols on the network. This movement is attributed mostly to increased scalability in application deployment, reducing the overall cost of network resources. Without the identification of critical applications, networks run the risk of becoming overwhelmed by lower-priority, less important protocols, reducing resources
available to higher priority applications. Deep packet inspection and statistical characteristic profiling have been used to identify various application flows, however there is an apparent disconnect between the two. Whilst both produce fairly accurate results, this paper aims to increase the accuracy of these systems by marrying the two into a single classifier using
artificial neural networks.
Whilst many traffic profiling systems examine the full network flow post-termination, this paper proposes a methodology for utilizing the unique characteristics of network traffic flows which distinguish various applications at the beginning of the flow, in real-time, allowing early identification and thus effective control of a flow within the first few packet exchanges.
Download count: 1377
How to get this paper:
PDF copy of this paper is free to download. You may distribute this copy providing you cite this page as the source.