A Method of Estimating Statistical Characteristics of Network Traffic After Its Decorrelation

The paper considers a method of estimating statistical characteristics of a sequence of time intervals subjected to decorrelation and characterizing the time between packets in telecommunication traffic. The traffic of modern applications has pronounced correlation properties, and the classical queueing theory assumes the absence of correlation. The paper proposes the implementation of decorrelation of the initial correlated sequence of time intervals and the subsequent estimation of their statistical characteristics based on cumulant analysis.