Skip to Main Content
IP networks are more and more used to transport real-time services. A crucial factor for real-time services is the end-to-end delay experienced by the application. The contribution resulting from the queuing delay induced by the network nodes is the most difficult to calculate. Under assumptions of real-time traffic behaviors and network nodes behaviors, the transportation of real-time traffic is modeled as a system of N statistically independent M/G/1 nodes. Then an approximate method is proposed to calculating (1-P)-quantile of queuing delay for real-time services, which is based on the dominant pole concept, and makes use of the fact that the total queuing delay is the sum of individual queuing delays that have approximately an exponential tail distribution function. Results are shown indicating an improvement up to 30% over the deterministic method. The formulae developed in this paper are expected to be useful in network dimensioning, setting network performance requirements and admission control mechanisms.
Date of Conference: 22-23 May 2010