Sample Based Estimation of Network Traffic Flow Characteristics.
Lili Yang · Deep Blue (University of Michigan) · 2008
Understanding the characteristics of traffic flows is crucial for allocating the necessary resources (bandwidth) to accommodate users demand. In this dissertation research, the problem of nonparametric and Moment-Based estimations of network flow characteristics based on sampled flow data from single-stage Bernoulli sampling, and two-stage sampling will be addressed. An Expectation-maximization (EM) algorithm is used for the flow length distribution, which in addition provides an estimate for the number of active flows. The estimation of the flow sizes (in bytes) is accomplished through a regression model. A variation of this approach, particularly suited for mixture distributions that appear in real network traces, is also considered. Lastly, estimation of traffic characteristic across the network and sampling allocation problem is studied. The proposed approaches are illustrated and compared on a number of synthetic and real data sets.