Choosing strategic locations to optimally place gateways prior to network deployment in wireless mesh networks (WMNs) can alleviate a number of performance related problems; it can also lead to better handling of network scalability. Existing solutions that address the optimal gateway placement problem differ mainly in terms of the set of constraints that the placed gateways has to satisfy; the resulting placements influence, differently, the network quality of service (QoS). In this paper, we study the WMN topology design and we propose a clustering based gateway placement algorithm (CBGPA) that guarantees end-to-end bounded delay communications with a good handling of network scalability. We show, via a case study, that CBGPA is constraints-independent algorithm that can effectively be coupled with a WMN design model; for that, we propose a multi-objective optimization model to design WMNs topologies from scratch. The two objectives of deployment cost and average congestion of gateways are simultaneously optimized in the model. The optimization model proposed is solved using a nature inspired meta-heuristic algorithm coupled with CBGPA, which provides the network operator with a set of bounded-delay tradeoff solutions. A comparative experimental study, using large size networks (up to 169 nodes) and different key parameter settings is conducted to show the effectiveness of CBGPA and to evaluate the performance of the proposed model.