On flow level modeling of multi-cell wireless networks

In order to meet increasing traffic demands, future generations of cellular networks are characterized by decreasing cell sizes at full frequency reuse. Due to inevitable inter-cell interference, load conditions in neighboring cells can no longer be considered independent. This paper provides a flow level modeling framework for cellular networks, where the coupling of flow level dynamics due to intercell interference is specifically taken into account. Since an adequate queueing model renders analytically intractable, we review different methods from the literature to bound and approximate the stationary behavior of network performance measures. Numerical investigations of a typical wireless scenario reveal, that in high and low load regimes first as well as second order bounds may be quite loose, depending on the type of bound. Especially for design of network optimization algorithms, bounds do not appear to suitably reflect network performance, and approximation techniques must be considered instead. In this regard, a suitable tradeoff between computational complexity and accuracy over the whole traffic range is provided by a model based on the notion of average interference.