Vasilios A. Siris,
Institute of Computer Science (ICS), FORTH
ICS-FORTH Technical Report No. 299, December 2001.
Congestion pricing has been identified as a flexible mechanism for efficient and robust resource control in fixed wireline networks. In this report we investigate the application of congestion pricing in Wideband CDMA (Code Division Multiple Access) networks. We begin by discussing the problem of quantifying resource usage for both the uplink and the downlink in CDMA networks. In the uplink, resource usage is an increasing function of the product of the transmission rate and the signal quality, given by the target bit-energy-to-noise-density ratio. In the downlink, resource usage is given by the transmission power from the base station. Based on these results, we propose and analyze a congestion pricing framework for elastic traffic in WCDMA networks, identifying how the properties of wireless networks affect the application of congestion pricing models. Our framework seeks to exploit the joint control of the transmission rate and signal quality in order to achieve, in a distributed and decentralized manner, economically efficient utilization of wireless network resources. The framework incorporates the congestion charge for shared resources in the wireless and wireline network, as well as the cost of battery power consumption at the mobile hosts, and is extended to the case of hybrid code and time division scheduling. For elastic traffic we show that the net utility maximization problem can be decomposed into two simpler problems: one involving the selection of the optimal target bit-energy-to-noise-density ratio, and one involving the selection of the optimal transmission rate. This result can simplify the integration of rate control at the CDMA and the transport layer. Whereas the above considered resource control on timescales larger than those of fast closed-loop power control, our final investigation involves the application of congestion pricing for achieving fair, efficient, and robust power control. Keywords: wireless data networks, resource usage, rate control, power control, elastic traffic, utility, social welfare