Analysis of Metrics for Routing Optimization in Wireless Mesh Networks Daniel Pereira, Alicia Triviño-Cabrera, and Marilia Curado Abstract – Wireless Mesh Networks are becoming a reality in today’s communication in wireless and mobile environments. Routing is a key component in Wireless Mesh Networks and the choice of the adequate routing metrics has been widely discussed. In this paper, a thorough analysis of the impact of routing metrics on traffic performance is carried out. Moreover, a new interference aware routing metric is presented and evaluated by simulation. This metric, resulting from a composition of metrics that, separately, deal with delay and packet loss takes into account application requirements aiming at providing different levels of Quality of Service. Our results show that the proposed metric improves the loss rate and fulfils the application requirements while maintaining the end-to-end delay values of the compared metrics. Index Terms— Mesh networks, routing metrics, interference I. INTRODUCTION Wireless networks are used in the most diverse and varied ways, not only because of the increasing number of mobile devices, but also because they represent a good investment, due to the low prices of wireless hardware. However, the use of wireless networks brings some technological problems as current protocols are highly customized for wired networks. 1 Wireless Mesh Networks (WMN) are included in the set of wireless networks that suffer from that performance hit. WMN bring, however, benefits. They are networks where each node is interconnected with several other nodes and can be either a host or a router, forwarding data as needed. WMN complement this scenario with the inclusion of a backbone, usually static, that provides connectivity to mobile nodes with a wider set of characteristics. There 1 Daniel Pereira is with the Department of Informatics Engineering of the University of Coimbra, Portugal, (phone: +351-239-790018; dvieira@student.dei.uc.pt). Marilia Curado is with the Department of Informatics Engineering of the University of Coimbra, Portugal, (phone: +351-239-790000; fax: +351-239-701266; marilia@ dei.uc.pt). Alicia Triviño-Cabrera is with the University of Malaga, Spain (atc@uma.es). is also an economic benefit, as these networks are self- organized and can be self-healing [1]. The quality of the paths of multi-hop wireless networks is severely affected by the interference between multiple simultaneous transmissions [2]. Shortest path algorithms based on the hop count metric are not suited for multi-radio scenarios because the path selection does not guarantee that the links are on different channels and therefore the total available capacity is not completely used due to interference. In order to address this issue, a novel routing metric which quantifies the quality of wireless paths according to losses, delay and interference is presented and evaluated by simulation. This paper is organised as follows: Section 2 describes the main proposals for routing metrics in WMN; Section 3 presents the proposed quality and interference aware routing metric; the comparison of the proposed metric with existing metrics is performed be simulation and the results are presented and discussed in Section 4; finally, Section 5 describes the main conclusions of the paper. II. ROUTING METRICS FOR MESH NETWORKS In wireless networks, sources may know multiple routes for a destination. Therefore, multi-hop communication in WMN requires routing metrics that represent the quality of the paths in order to maximise network use and to enhance traffic performance. The hop count metric, commonly employed in wired networks, is not suited for WMN since it is oblivious to important characteristics associated with the wireless technology, namely, variable loss rate due to the change of channel properties over time and interference caused by broadcast communication in the wireless links. In this context, routing metrics for WMN should allow the selection of paths that minimise loss rate, interference and maximise throughput [3]. The two main approaches to solve this problem are to develop a routing metric that represents these three aspects, or to conceive a routing algorithm that computes multi-constrained paths according to all the metrics. In both situations, there is the need to