Architectures and deployment strategies for wireless mesh networks

Abstract

Nowadays the development of the next-generation wireless systems (e.g., the fourthgeneration (4G) mobile cellular systems, IEEE 802.lln, etc.) aims to provide high data rates in excess of 1 Gbps. Thanks to its capability of enhancing coverage with low transmission power, wireless mesh networks (WMNs) play a significant role in supporting ubiquitous broadband access [1]- [10]. Fig. 2.1 illustrates a multi-hop wireless mesh network, where only the central gateway G has a wireline connection to the Internet and other nodes (like node S) access to the central gateway via a multi-hop wireless communication. Each node in the WMN should operate not only as a client but also a relay, i.e., forwarding data to and from the Internet-connected central gateway on behalf of other neighboring nodes. The main difference between ad hoc networks and wireless mesh networks is the traffic pattern [2], as shown in Fig. 2.2. In a WMN, there will exist a central gateway and most traffic is either to/from the central gateway as shown in Fig. 2.2(a). In an ad hoc network, however, traffic flows are arbitrary between pairs of nodes, such as the flow between nodes S1 and D1 in Fig. 2.2(b). In general, the advantages of wireless mesh networking technology can be summarized into five folds. First, WMN can be rapidly deployed in a large-scale area with a minimal cabling engineering work so as to lower the infrastructure and deployment costs [1]- [5]. Second, mesh networking technology can combat shadowing and severe path loss to extend service coverage area. Third, by means of short range communications, WMN can improve transmission rate and then energy efficiency. In addition, the same frequency channel can be reused spatially by two links at a shorter distance. Fourth, due to multiple paths for each node, an appealing feature of WMNs is its robustness [9], [10]. If some nodes fail (like node B in Fig. 2.3), the mesh network can continue operating by forwarding data traffic via the alternative nodes. Fifth, WMN can concurrently support a variety of wireless radio access technologies, thereby providing the flexibility to integrate different radio access networks [6]- [8]. Fig. 2.4 shows an example of integrated wireless mesh network, where 802.16 (WiMAX), 802.11 (WiFi), and 802.15 (Bluetooth and Zigbee) technologies are used for the wireless metropolitan area network (WMAN), the wireless local area network (WLAN), and the wireless personal area network (WPAN), respectively. However, when the coverage area increases to serve more users, multi-hop networking suffers from the scalability issue [10]. This is because in the multi-hop WMNs throughput enhancement and coverage extension are two contradictory goals. On one hand, the multi-hop communications can extend the coverage area to lower the total infrastructure cost. On the other hand, as the number of hops increases, the repeatedly relayed traffic will exhaust the radio resource. In the meanwhile, the throughput will sharply degrade due to the increase of collisions from a large number of users. Therefore, it becomes an important and challenging issue to design a scalable wireless mesh network, so that the coverage of a WMN can be extended without sacrificing the system overall throughput. In this chapter, we first discuss the major architectures of WMNs and briefly overview the existing mesh networking technologies, including the IEEE 802.11s and IEEE 802.16 systems. Then, we address the scalability issue of the WMN from a network deployment perspective. We introduce two scalable-WMN deployment strategies for the dense-urban coverage and wide-area coverage scenarios as shown in Figs. 2.5 and 2.6 ( [11, 12]). First, the cluster-based wireless mesh network for the dense-urban area is shown in Fig. 2.5. In this WMN, several adjacent access points (APs) form a cluster and are connected to the Internet through the same switch/router. In each cluster, only the central access point AP0 connects to the Internet through the wires. Other APs are interconnected by wireless links. By doing so, the network deployment in the urban area becomes easier because the cabling engineering work is reduced. Second, a scalable multi-channel ring-based WMN for wide-area coverage is shown in Fig. 2.6, where the central gateway and stationary mesh nodes in the cell form a multi-hop WMN. Note that the mesh cell is divided into several rings allocated with different channels. In the same ring, the mesh nodes can follow the legacy IEEE 802.11 medium access control (MAC) protocol to share the radio medium. Besides, mesh nodes in the inner rings will relay data for nodes in the outer rings toward the central gateway. Based on this mesh cell architecture, the service coverage of the central gatewary/AP can be effectively extended with a lower cost. We will also investigate the optimal tradeoff between capacity and coverage for these two scalable WMNs. Most traditional wireless mesh networks are not scalable to the coverage area because the user throughput is not guaranteed due to the increase of collisions. By contrast, theWMNs shown in Figs. 2.5 and 2.6 are more scalable in terms of coverage because frequency planning with multiple channels can be easily applied in this architecture to resolve the contention issue. Thus the throughput can be ensured by properly determining the deployment parameters. We will apply the mixed-integer nonlinear programming (MINLP) optimization approach to determine the optimal deployment parameters, aiming to maximize the capacity and coverage simultaneously. The rest of this chapter is organized as follows. Section 2.2 presents the major network architectures for WMNs. Sections 2.3 and 2.4 discuss the mesh networking technologies in the IEEE 802.11s and IEEE 802.16 systems, respectively. Section 2.5 describes the proposed scalable wireless mesh networks for the dense-urban coverage and the wide area coverage. In addition, we apply the optimization approach to determine the optimal deployment parameters, aiming at maximizing the coverage and capacity. At last, concluding remarks are given. © 2007 Springer Science+Business Media, LLC. All rights reserved.