Random early detection (RED) controlled loss (i.e., discarding data packets) is determined as a function of change in processing gain assigned by a resource management system in a data network having a communications link between first and second network nodes. Rather than triggering RED controlled loss as a function of buffer levels, triggering is determined as a function of change in processing gain caused by, for example, a change in code rate, modulation technique, error (e.g., bit error rate or frame error rate), signal-to-noise ratio (SNR) or carrier-to-interference (C/I) level, or a number of traffic code channels or TDMA slots assigned to the nodes. In a wireless data network, this technique may be deployed in a base station or access terminal. A tight coupling between the physical layer and link layer is provided using this technique.

To select a transmission mode to use for a data transmission via a MIMO channel from station A to station B, station A obtains channel information used for spatial processing and determines the age of this information. Station A selects one of multiple transmission modes based on the age of the channel information and possibly other information (e.g., the fading characteristic of the MIMO channel). To select rate(s) to use for the data transmission, station A obtains channel state information (CSI) indicative of the received signal quality for the MIMO channel, e.g., received SNRs or "initial" rates. Station A determines the age of the CSI and selects one or more "final" rates based on the CSI, the age of the CSI, the selected transmission mode, and possibly other information. Station A processes data in accordance with the selected transmission mode and final rate(s) and transmits the processed data to station B.