NETWORK DESIGN

Network requirements and design



Network requirements and design

Introduction

Wireless computing and communications devices have permeated every aspect of business and personal lives. With the increasing interest and developments in wireless technologies, wireless mesh networks (WMNs) have started to emerge as the most promising model for integrating different wireless networks and for providing broadband access, availability, and coverage to various wireless clients. WMNs are usually touted in the literature as a concept network with a rich set of desirable features and capabilities. They are typically capable of self-organizing, self-healing, self-configuring, and self-maintaining, while providing high availability, resilience, and fault-tolerance low cost (Akyildiz, 2009, 66).

Wireless mesh networks have their own features and characteristics that distinguish them from Mobile Ad hoc NETworks (MANETs), wireless sensor networks (WSNs), and wireless local area networks (LANs). For instance, the mobility and power consumption is not much of a concern in WMNs because the mesh routers usually do not move, and they have power supplies. Only the mesh clients can move partially and may have limited power supplies. In addition, WMNs can have clients and routers with multiple radios/channels. Routing with multiple channels can provide better coverage and throughput to boost the performance of voice over Internet Protocol (VoIP), multimedia teleconferencing, and video-on-demand applications that require higher bandwidth and lower end-to-end delays.

Within the last decade, there has been a lot of research on WMNs. This research has mainly focused on routing and medium access control (MAC) protocols, channel assignment, QoS provisioning, and interference avoidance. In these research efforts, the performance of the proposed protocols on the aforementioned topics has been widely tested in network simulators, such as NS-2, NS-3, QualNet (Scalable Network Technologies, Inc., Los Angeles, CA, USA), Glomosim , OPNET (OPNET Technologies, Inc., Bethesda, MD, USA) , and JSim .

Although this approach works to some extent, the performance evaluations may not be applicable to real-life wireless networks because of several unrealistic assumptions within the simulations, and the inability of the simulators to accurately reflect the behavior of the wireless environment . Inaccuracy of lower-layer modelling, especially of the physical layer, such as propagation environment, interface characteristics, and path loss parameters, are elaborated in and . The authors further noted that granularity in simulations is at packet-level, whereas the theoretical analysis these simulations are based on use bit or symbol-level derivations. Several studies, such as note the deficiency of unit disk graph1 assumption to model the topology of the wireless networks. includes transceiver properties, propagation model, radio link model, interference model, and energy consumption parameters as part of the unrealistic assumptions in wireless simulations (ZhangY, 2007, 63).

Inadequacy of simplistic mobility model and radio wave propagation characteristics in wireless simulations are explained in. Imprecision of multichannel interference modelling is discussed in. Inaccuracy of neighbourhood discovery and radio link symmetry in simulations is described in. Negative impact of incorrect physical layer assumptions, such as signal reception, path loss, fading, interference, noise computation, and preamble length, lead to false performance evaluation at higher ...