Dynamic and robust capacity allocation in wireless networks carrying heterogeneous traffics

Abstract

We consider wireless networks carrying heterogeneous traffics. We assume that the characteristics of the traffic classes are time varying: at each point in time, a traffic class is described by a process randomly selected from within a set of non parametrically defined stochastic processes.For the wireless environment described above, we compare two multiplexing transmission techniques: a dynamic Time-Division-Based and a dynamic CDMA-Based techniques, in terms of a number of pertinent performance metrics. The dynamics of both techniques manifest themselves as dynamic capacity allocations to the various traffic classes and are implemented via the deployment of a Robust Traffic Monitoring Algorithm (RTMA). The performance metrics studied include delays, traffic rejection rates and wasted capacity rates. Qualitatively speaking, a major controlling factor in the choice among the Time-Division-Based and the CDMA-Based techniques is the speed of the transmission channel: high speeds favor CDMA-Based transmissions.