**2.1 (Mobile) Ad-Hoc Networks**

An *Ad-Hoc Network* is defined as a network where all nodes communicate with one another on an ad-hoc basis without a central base station [7]. While sometimes the term is used in literature to denote a Wireless Multi-Hop Network, we use it here to denote a wireless network that does not differentiate between client nodes and dedicated routing nodes. The typical node is a somewhat powerful device such as a (ruggedized) laptop, smartphone, first-responder communication device, or a device that is integrated in a vehicle, all of which are possibly mobile (see **Figure 1**). In this case, the network is called a *Mobile Ad-Hoc NETwork* (*MANET*). These networks were originally developed for military use to enable troop communications in areas where no communications infrastructure was previously deployed. *MANETs* are also envisioned for emergency and disaster networking where the borders between *MANETs* and *WMNs* (see below) are flowing [8].

Several "classical" routing algorithms have been developed for *MANETs*, the most prominent being:

DSR: Dynamic Source Routing Protocol was developed in 1994 by David B. Johnson [9]. It is a *reactive* protocol which means that the protocol only builds routes on-demand. This is advantageous in highly volatile networks where it makes no sense to invest routing overhead to build and maintain routes that might go stale before they are used. On the other hand, traffic incurs a routesetup delay because the route is not built in advance. DSR is being standardized by the IETF [10],

DSDV: Destination-Sequenced Distance-Vector Protocol was developed by Charles Perkins and Pravin Bhagwat in 1994 [11]. It is a *pro-active* routing protocol, which means that routes are built in advance. This avoids the routesetup delay incurred by reactive protocols but, in highly volatile networks, a lot of routing overhead might be spent to set up and maintain routes that break before they can be used,

*Ant Algorithms for Routing in Wireless Multi-Hop Networks DOI: http://dx.doi.org/10.5772/intechopen.99682*

### **Figure 1.**

*MANET architecture. Nodes connect on an ad-hoc basis without dedicated routers or base stations.*

AODV: Ad hoc On-Demand Distance Vector Routing [12] developed 1999 by Charles Perkins and Elizabeth Royer which is a reactive routing protocol, and OLSR: Optimized Link State Routing Protocol [13] which was developed by Jacquet et al. in 2001. It is a proactive routing protocol and has been standardized by the IETF as experimental RFC 3626 [14].

### **2.2 Sensor Networks**

A *Sensor Network* (or *Wireless Sensor Network*, *WSN*) is a wireless network that consists of many sensor nodes which are spatially deployed to cooperatively monitor physical or environmental conditions (see **Figure 2**). *WSNs* were initially developed for the military for applications such as battlefield surveillance. However, *WSNs* are now used in civilian application areas like environmental monitoring [15–17]. In contrast to *MANETs* sensor nodes are typically tiny (so-called "motes") and deployed densely in huge numbers, often on inaccessible terrain. Therefore, sensor network protocols must be self-organizing.

One of the major problems in sensor networks is the limited power as each sensor only has a small battery. While some sensors use technologies like solar cells to refresh their batteries, routing protocols for sensor networks typically try to optimize for power efficiency [18]. In *WSNs*, typically the sink initializes routing by issuing a query for measurement data. Sensor nodes answer the query by sending their data back to the sink. To save energy, data may be aggregated along the way (*data-centric routing*). To facilitate this, the sensor field is often divided into clusters or subnets. All nodes of a cluster first send their data to the respective cluster head, which then processes and routes the data to the sink. If the sensors are equipped with location finding devices like GPS (Global Positioning System), knowledge of the position can be used to ease cluster formation or perform geo-routing. Well known non-ant routing protocols for *WSNs* include:

Gossiping is an early approach derived from flooding [19], SPIN: Sensor Protocols for Information via Negotiation is a family of protocols based on data-centric routing, developed by Heinzelman et al. in 1999 [20],

### **Figure 2.**

*WSN architecture. Sensor nodes are deployed in a sensor field and deliver their measurements through a sink to the Internet.*


### **2.3 Wireless Mesh Networks**

A *Wireless Mesh Network* (*WMN*) is a wireless networking architecture in which nodes are connected via a *wireless backbone* [26]. In contrast to *MANETs*, though, the wireless backbone in a *WMN* is typically fixed. Iow. a *WMN* consists of nonmoving wireless mesh router nodes which constitute the wireless backbone and (potentially mobile) client nodes (see **Figure 3**). Router nodes can be mesh routers only (so-called Mesh Points, MP [27]) or act as combined *WLAN* Mesh Access Points (MAPs)/routers.

*WMNs* can be used as access networks to the Internet, where one or several Mesh Portal Points (MPP) connect the mesh to the Internet. They can also be used in disaster areas, for emergency response teams, and for the military. The boundary towards *MANETs* is somewhat fluid in these cases [8]. Consider, for example, a rescue operation where wireless routers are installed on top of firetrucks. When the firetrucks arrive at the scene, they stop and provide the wireless backbone for the firefighters' communication devices. This scenario can be seen as a *WMN* as well as as a *MANET*.

Even with a stationary wireless backbone, the characteristics of wireless channels and the interaction of the MAC layer with the higher layers in the network

**Figure 3.**

*WMN architecture. Client nodes connect via a wireless backbone. MP, Mesh Point; MAP, Mesh Access Point; MPP, Mesh Portal Point.*

stack make routing in *WMNs* a hard problem. Wireless links vary over time and problems like the hidden node problem or the exposed node problem influence routing algorithm's performance. Therefore, a "wireless-aware" routing algorithm is necessary.

*WMNs* have slightly different constraints and pose different problems than *MANETs* and sensor networks. For example, the power constraint which is very prominent in sensor networks typically does not exist in *WMNs* [8]. When *WMNs* are used as access networks to the Internet "normal internet traffic" has to be assumed. This means application traffic like streaming, web browsing, VoIP (Voice over IP) and video conference traffic, or email, which use the standard TCP (or UDP) protocol stack. Routing algorithms for *WMNs* have mostly been adapted from the (Mobile) Ad-Hoc Networking Community.
