Estimation of the Probability of Buffer Overflow for Self-Similar Traffic

Abstract

In recent years, a large number of studies have been carried out on the traffic of real telecommunication systems. It has been found that traffic has a number of specific properties that distinguish it from regular traffic. Namely: it has the properties of self-similarity, long-term dependence and distribution of the magnitude of the load coming from one source, has heavy tails. New traffic models were built that had the specified properties. The work studies such models, methods of traffic modeling. Estimates of the probability of exceeding the buffer of self-similar traffic are obtained.