Interval Method of Traffic Analysis in Multiservice Access Networks

Abstract

The article analyzes queuing systems’ stationary, ordinary recurrent streams of applications. The effect of correlation between the incoming flows of requests for queue sizes is determined. The article shows that the queue formation process in systems with a batched stream is cyclic. This process is divided into two sub-cycles: active and passive. Due to cycle analysis the generalization of the well-known Khinchin-Pollaczek formula is found. This formula was applicable only for Poisson queue flows before. The determined effect of correlations on the queue’s size is shown. The article proves that the maximum value of the covariation function’s integral of the applications’ number on intervals of the application’s service fully determines the middle size of queue into primary queuing systems with established load factor.