**3. The study of optimal number of radio interfaces**

In this section several simulations were created to find the optimal number of radio interfaces for static WMN. In this study we focus only to one problem - to find the optimal number of radio interfaces for different conditions therefore, for channel assignment we used simple CCA approach (section 5.1).

Nowadays the availability of the cheap off-the-shelf commodity hardware also makes multiradio solutions economically attractive. This condition provides the using much more radio interfaces per node, which shows the investigating of optimal number of interfaces as a reasonable argument.

We have included in our simulations several QoS parameters, data flows, number of nodes and network topologies to find the optimum number of radio interfaces for services which required the real time transmission (e.g. video conference).

#### **3.1. Simulation environment**

82 Wireless Mesh Networks – Efficient Link Scheduling, Channel Assignment and Network Planning Strategies

parameters.

**2. Related work** 

put).

CCA approach (section 5.1).

reasonable argument.

parameters. In the second part of the chapter, the new proposed centralized channel assignment concept called First Random Channel Assignment algorithm (FRCA) is compared with two other channel assignment techniques (CCA, LACA) by the same QoS

The rest of this chapter is organized as follows. In section 2, the related work is summarized. In section 3, the methods and simulation results to find the optimal number of radio interfaces per node are introduced. In section 4 the mathematical background and graph based mathematical model is described and in the next section different types of channel assignment methods based on links load are analyzed. Section 6 concludes the chapter.

There exist a large number of studies which address the channel assignment problem in wireless mesh networks. Several works have proposed MAC protocols for utilizing multiple channels (So & Vaidya, 2004, Gong & Midkiff, 2005), but these multi-channel protocols require changes to existing standards and therefore cannot be deployed by using existing hardware. In (Adya et al., 2004) was proposed a link-layer solution for transmitting data over multiple radio interfaces, but this approach is designed for scenario where the number of radio interfaces is equal to the number of channels. In (Gupta & Kumar, 2000) the performance of multi-channel ad-hoc networks was studied, where each channel was assigned to an interface. In (Draves et al., 2004) several methods for increasing the performance in single-channel per interface were proposed. The most studies is focused only to one problem to find the efficient channel assignment method, but did not suggest the optimal number of radio interfaces per node. In (Husnain et al., 2004) were compared different static centralized algorithms, but for evaluation of optimal number of radio interfaces was used only one parameter - total interference (number of links in conflict graph). (Raniwala et al., 2004) proposed centralized channel assignment and routing method, where results about number of radio interfaces were shown but only for network cross-section goodput. In (Chi Moon Oh et al., 2008) the study of optimal number of radio interfaces was created but only for grid network, using simple channel assignment method and for one QoS parameter (through-

In this section several simulations were created to find the optimal number of radio interfaces for static WMN. In this study we focus only to one problem - to find the optimal number of radio interfaces for different conditions therefore, for channel assignment we used simple

Nowadays the availability of the cheap off-the-shelf commodity hardware also makes multiradio solutions economically attractive. This condition provides the using much more radio interfaces per node, which shows the investigating of optimal number of interfaces as a

**3. The study of optimal number of radio interfaces** 

A simulation WMN model was developed in NS-2 network simulator, with additional function to support multi-channel and multi-interface solution (Calvo & Campo, 2007). Each mesh node used the number of interfaces between 1 to 8 and the same number of channels. Two different network topologies were created. The first one was grid topology, which consisted of 25 static wireless mesh nodes placed in an area of 1000 x 1000 meters. Transmission range for each node was set to 200 meters (Fig.4a). The second topology consists of 25 nodes, which were randomly placed in an area of 1000 x 1000 meters (Fig.4b). For simulation evaluations, ten random topologies and computed average values of chosen QoS parameters were studied. We have used the WMN with 25 nodes, because of the typical number of mesh nodes in WMN (25 to 30) (Skalli et al., 2006). For traffic generation, 5 CBR (Constant Bit Rate) flows were used and the packet size was set to 512 bytes. The same radio default parameters as in (ns-2, 2008) were used, except that we set the channel data rate to 11 Mbit/s. Simulation parameters are summarized in Table 1.


**Table 1.** Simulation parameters

**Figure 4.** Grid (a) and random (b) topology of static WMN created in NS-2 simulator
