Modeling End-to-End Delay Using Pareto Distribution

Delay is one of the network performance parameters that are often measured using passive or active techniques along with packet loss, bandwidth, etc. If used appropriately, these parameters can indicate performance status of the network, and they can be used in fault and performance management, network provisioning, traffic engineering, and performance prediction. However, it is difficult to extract sufficient information from original measurement data to derive precise results. In this paper, we show that the PDF (probability distribution function) of delay data can indicate network load situation. We also show that Pareto distribution can model end-to-end delay appropriately in a statistical manner and demonstrate some statistical characteristics of end-to-end delay using Pareto distribution.