In wireless mesh networks (WMNs), gateway nodes may become a severe bottleneck for Internet flows. Indeed, if traffic is routed in the mesh without considering traffic distribution, as well as link capacities, some gateways or intermediate mesh routers may rapidly get overloaded due to unevenly utilization of network resources. To address this unevenly utilization of network resources. To address this issue, in this paper we firstly develop a multi-class queuing network model to analyze feasible throughput allocations in heterogeneous WMNs, as well as to predict the residual capacity of network paths. Guided by our analysis, we design a Capacity-Aware Route Selection algorithm (CARS), which allocates network paths to downstream and upstream Internet flows so as to ensure a more balanced utilization of wireless network resources and gateways' Internet connections. Through simulations in a number of different network scenarios we show that CARS significantly outperforms shortest path routing using routing metrics that capture only inter-flow interference.