Performance Analysis of A Hybrid Scheduling Scheme Under Heterogeneous MMPP and Self-Similar Traffic

This paper presents an analytical model for the hybrid PQ-GPS scheduling scheme that integrates priority queueing (PQ) and generalized processor sharing (GPS) in a hierarchical manner for the provisioning of differentiated quality-of-service. To capture the heterogeneous traffic properties in multi-service networks, the Markov-modulated Poisson process (MMPP) and the self-similar fractional Brownian motion (fBm) are used to model short-range dependent (SRD) and long-range dependent (LRD) traffic, respectively. We propose an efficient flow-decomposition method and derive the queue length distributions and loss probabilities of individual traffic flows in the hybrid scheduling system. The comparison between analytical and simulation results validates the accuracy of the analytical model. To illustrate its application, we employ this model as an efficient tool to study the resource allocation issue under the constraint of loss probability.