Tail Probabilities of the Delay Under a Reservation-Based Scheduling Mechanism

Abstract

We study the tail probabilities of the customer delay for a buffer operating under a reservation-based scheduling discipline known as R-scheduling. Previous numerical work on this model has led to a hypothesized meta-model, which was tested using simulations but not proven analytically. In this paper, we prove the correctness of this meta-model analytically, and extend it to more general arrival processes. The results are also compared to simulations for several example scenarios.