Design of scalable and efficient multi-radio wireless networks

Djohara Benyamina, Abdelhakim Hafid, Michel Gendreau
2011 Wireless networks  
A proper design of Wireless Mesh Networks (WMNs) is a fundamental task that should be addressed carefully to allow the deployment of scalable and efficient networks. Specifically, choosing strategic locations to optimally place gateways prior to network deployment can alleviate a number of performance/scalability related problems. In this paper, we first, propose a novel Clustering Based Gateway Placement Algorithm (CBGPA) to effectively select the locations of gateways. Existing solutions for
more » ... ptimal gateway placement using clustering approaches are tree-based and therefore are inherently less reliable since a tree topology uses a smaller number of links. Independently from the tree structure, CBGPA strategically places the gateways to serve as many routers as possible that are within a bounded number of hops. Next, we devise a new multiobjective optimization approach that models WMN topologies from scratch. The three objectives of deployment cost, network throughput and average congestion of gateways are simultaneously optimized using a nature inspired meta-heuristic algorithm coupled with CBGPA. This provides the network operator with a set of bounded-delay trade-off solutions. Comparative simulation studies with different key parameter settings are conducted to show the effectiveness of CBGPA and to evaluate the performance of the proposed model. Index Terms-Wireless mesh networks, Network design and optimization, Clustering, Constraint satisfaction, Scalability. goal when designing multi-radio wireless networks since the impact of inter-path (or co-channel) interferences on network performance is reduced. Problem addressed. In this study, we address the problem of WMN design by exploiting the tradeoffs among network deployment cost, network throughput, gateway congestion level, the number of communication hops between sources and gateways, and user coverage. Indeed, minimizing the cost requires stingy resources utilization (deploying fewer routers and/or gateways) which impacts the network performance. With few routers deployed, the traffic is routed on longer paths to get to its destination, thus increasing communication delays. With few gateways deployed, congestion may happen (since all traffic traverse gateways to and from Internet) impacting network throughput. Conversely, deploying more resources (higher deployment cost) helps providing shorter paths and less congested gateways; however, this may cause high interference levels and thus degrade network performance. In fact, optimizing one of these criteria will affect/undermine other criteria; therefore, it is difficult, if not impractical, to have a solution that is optimal in all criteria.
doi:10.1007/s11276-011-0388-x fatcat:qrwz2inuqjbghaiqrdnkpapcqu