The article presents an approach to predicting changes in the dynamics of events, on the example of traffic analysis, using statistical data. Perfomed analisys of traffic sniffer systems as queueing system. Presented the relevance of the forecasting traffic changes. The pattern matching algorithm allows you to build a forecast, both for regular dynamics, and for unexpected bursts. The prediction algorithm allows to improve the work of traffic analysis systems for periods of high load.
Keywords: Network monitoring, traffic analysis, forecasting, queuing system, random events, service requests