Adaptive allocation of CDMA resources for network-level QoS assurances

In a Code Division Multiple Access (CDMA) network, multiple mobile hosts (MHs) can simultaneously transmit over the wireless channel by using different codes. To assure an acceptable quality of service for all users' flows, the network usually tunes the transmit powers of all MHs to achieve a certain level of signal strength as compared to the noise and the interference (SINR) for each user. The traditional assumption in power control schemes is that the SINR requirement is statically determined for each user flow. In contrast, in this paper, we propose a scheme that dynamically adapts the SINR requirements of user flows based on its quality of service (QoS) requirements and the conditions of the wireless channel between the MHs and the base station. As a result of this adaptation, we show that network-level QoS measures such as fraction of packets meeting their delay requirements and energy consumed per packet transmission are significantly better than in a scheme that statically fixes the SINR requirements. Our scheme uses a simple table-driven approach for optimally selecting the target SINR requirement for each MH at run time. The entries in the table are computed offline using a dynamic programming algorithm with the objective of maximizing a profit function that balances the need for meeting the network-level QoS requirements and the cost of using a particular target SINR for a given transmission.