Analytical modeling of an integrated priority and WFQ scheduling scheme in multi-service networks

Abstract

The provisioning of differentiated Quality-of-Service (QoS) is a key objective in the design and deployment of convergent multi-service communication networks. Integrated traffic management schemes that combine various fundamental scheduling principles have been proposed for the QoS differentiation in cellular and IEEE 802.16/WiMAX-based broadband wireless networks. In this paper, cost-effective analytical performance models are devised for a new hybrid scheduling scheme based on the integration of the well-known Priority Queueing (PQ) and Weighted Fair Queueing (WFQ) scheduling principles. In these models, traffic arrival processes are represented, respectively, by non-bursty Poisson process and bursty-correlated Markov-Modulate Poisson Process (MMPP) and new closed-form expressions of main performance metrics are derived including system fairness, mean number of packets in each buffer, system throughput, mean queueing delay and packet loss probability of individual traffic flows. The accuracy of the analytical models is validated through extensive comparisons between analytical results and those obtained from simulation experiments. Finally, these models are adopted to investigate the impact of the weights of traffic flows served by WFQ and assess the adverse effects of the traffic burstiness and correlation on the performance of the integrated scheduling scheme.