Currently, we see the evolution of large scale community and metropolitan area networks based on inexpensive wireless local area network technology. We present the results of an experimental analysis which investigates the potential of quality of service routing mechanisms within this challenging environment. Our investigation is based on a model of a radio access network designed to cover a large city center by means of decentralized and distributed routers, which are tightly meshed. The workload is modeled to reflect the estimated usage patterns based on statistical data collection of user mobility and combined with synthetic traffic matrices. We present results for various routing strategies including shortest path routing, delay constrained routing as well as various multipath quality of service routing variants. Moreover, we investigate different traffic distributions. Our findings are, that multipath routing is able to enhance the utility of the network significantly.