Characterization of network traffic on a local area network

Abstract

In this thesis we analyze network traffic on a local area network in an academic environment. We analyze the trace based on transport-level and network-level applications. Networks conversations are studied using the newly developed packet-train model. Analysis is done in relation to the traffic breakdown, bulk data transfer, interactive applications, packet interarrival times, network utilization etc. We survey the available modeling techniques for network traffic and a characterization of the trace is done. Characterization using some linear and non-linear models is done including estimation of empirical parameters and finding the autocorrelation functions. The traffic is also analyzed in the frequency domain using fast Fourier transforms and digital filtering in the Fourier domain. A network traffic characterization tool is developed based on the above methodology. Index Items: Local-area networks, trace analysis, characterization, simulation, modeling, TCP/IP internetworks