ESTIMATION OF QOS CHARACTERISTICS OF SELF-SIMILAR TRAFFIC FOR THE WB/M/1/¥ QUEUING SYSTEM

Abstract

The work considers the problem of Quality of Service (QoS) characteristics estimation for self-similar traffic serviced by a queuing system of the WB/M/1/¥ type using the Weibull distribution. To solve this problem, a transformation of Laplace-Stieltjes is applied. The values of service quality characteristics, such as the average time of packets delay and the length of packets queue, have been obtained for self- similar traffic serviced by the WB/M/1/¥ queuing systems (QS). We obtained graphs of the dependence average waiting time W = W(l,H) of a request in the system and аverage queue length Q = Q (r,H) which allow us to provide practical guidance on the maximum value of the average waiting time W for queries that arrive at the WB/M/1/¥ QS. It can be stated when planning the sizes of the buffers for hardware and software facilities, it is necessary to plan them in the way that would guarantee the required QoS. The obtained results will allow the consideration of design features and structure of the network nodes at the stage of planning and during further operation of the NGN hardware and software facilities.