Estimating Pole Capacity in a Live HSDPA Network

Estimating the maximum capacity of a radio carrier is of key importance to dimension a mobile network cost-effectively. In practice, such an estimate is difficult to obtain since carrier capacity is influenced by many factors that are hard to measure and quantify. In this paper, an empirical performance model is proposed for estimating High-Speed Packet Downlink Access (HSDPA) pole capacity on a per-cell basis from statistics in the network management system. The proposed model is obtained by multiple regression analysis over measurements taken from cells in a live HSDPA network. Results show that, in spite of its simplicity, the model is able to predict the pole capacity on a per-cell basis fairly accurately.