End-to-end delay differentiation of IP traffic aggregates using priority queuing models

This article evaluates the use of priority queuing (PQ) models to achieve delay differentiation in networks operating under the class of services paradigm. Three PQ variants are considered: the proportional model, the additive model and a novel hybrid schema based on the upper time limit model. Through appropriate queuing and scheduling, delay differentiation is first considered for a node, and then extended to an end-to-end perspective. For each model, the corresponding delay differentiation bounds are determined and verified resorting to simulation.