Class-based downlink capacity estimation of a WCDMA network in a multiservice context

This paper presents a class-based downlink capacity estimation technique with several modifications to the general CBQ (Class-Based Queuing) principles to ensure its applicability in a mobile environment. Instead of the fixed bandwidth used in wired links, a simple analytical expression for the maximum available resource of a radio downlink is derived in the paper. The resource-sharing factor for each class is then defined as the total resource allocation for that class divided by the maximum available resource. Analytical result shows that the class-based downlink throughput can be expressed as the weighted sum of the resource-sharing factors of all classes. CBQ-based downlink capacity dimensioning, including estimations on the average number of connections for different classes and calculations for the estimated downlink throughput in times of congestion, is presented using the analytical model derived in this study. A simulative evaluation is provided to show the accuracy of the analytical model. When the downlink becomes congested, the simulative throughput is approaching to the estimated result and is restricted by this value. A simple CBQ-based call admission control (CAC) algorithm is developed in our simulation model.