Part I Introduction 1 Ad Hoc and Sensor Networks 1.1 The Future of Wireless Communication Recent emergence of affordable, portable wireless communication and computation devices and concomitant advances in the communication infrastructure have resulted in the rapid growth of mobile wireless networks. On one hand, this has led to the exponential growth of the cellular network, which is based on the combination of wired and wireless technolo- gies. Nowadays, the number of cellular network users is approaching two billion worldwide (expected at end 2005). Although the research and development efforts devoted to tradi- tional wireless networks are still considerable, the interest of the scientific and industrial community in the realm of telecommunications has recently shifted to more challenging scenarios in which a group of mobile units equipped with radio transceivers communicate without any fixed infrastructure. 1.1.1 Ad hoc networks Ad hoc networks are the ultimate frontier in wireless communication. This technology allows network nodes to communicate directly to each other using wireless transceivers (possibly along multihop paths) without the need for a fixed infrastructure. This is a very distinguishing feature of ad hoc networks with respect to more traditional wireless networks, such as cellular networks and wireless LAN, in which nodes (for instance, mobile telephone users) communicate with each other through base stations (wired radio antennae). Ad hoc networks are expected to revolutionize wireless communications in the next few years: by complementing more traditional network paradigms (Internet, cellular networks, satellite communications), they can be considered as the technological counterpart of the concept of ubiquitous computing. By exploiting ad hoc wireless technology, various portable devices (cellular phones, PDAs, laptops, pagers, and so on) and fixed equipment (base stations, wireless Internet access points, etc.) can be connected together, forming a sort of ‘global’, or ‘ubiquitous’, network. Application scenarios in which the adoption of ad hoc networking technologies might prove useful abound. For instance, consider the following situation. A terrible earthquake has Topology Control in Wireless Ad Hoc and Sensor Networks P. Santi  2005 John Wiley & Sons, Ltd 4 AD HOC AND SENSOR NETWORKS devastated the city of Futuria destroying, among other things, most of the communication infrastructure (wired phone lines, base stations for cellular networks, and so on). Several rescue teams (firefighters, police, medical teams, volunteers, and so on) are working on the disaster scene to save people from wreckage and to assist the injured. To provide a better assistance to the population, the efforts of the rescue teams should be coordinated. Clearly, a coordinate action can be achieved only if rescuers are able to communicate, both within their team (e.g. a policeman with other policemen) and with members of the other teams (e.g. a firefighter calling a doctor for assistance). With currently available technology, coordinating rescuers’ efforts when the fixed communication infrastructure is severely damaged is very difficult: even if team members are equipped with walkie-talkies or similar devices, when no access to the fixed infrastructure is available, only communication between nearby rescuers is possible. Thus, one of the priorities in present-day disaster management is to reinstall the communication infrastructure as quickly as possible, which is typically done by repairing the damaged structures and by deploying temporary communication equipment (e.g. vans equipped with a radio antenna). The situation would be considerably different if technologies based on ad hoc network- ing were available: by using fully decentralized, multihop wireless communication, even relatively distant rescuers would be able to communicate, provided there exist other team members in between them acting as communication relay. Since a disaster area is typically quite densely populated with rescuers, citywide (or even metropolitanwide) communication would be possible, allowing a successful coordination of the rescue efforts without the need for reestablishing the fixed communication infrastructure. The above-described example outlines the features of a typical ad hoc network applica- tion scenario: – Heterogeneous network: A typical ad hoc network is composed of heterogeneous devices. For instance, in the scenario described above, in general the various teams working on the disaster area are equipped with different types of devices: cell phones, PDAs, walkie-talkies, laptops, and so on. For a successful setup of the communication network, it is fundamental that these diverse types of devices be able to communicate with each other. – Mobility : In a typical ad hoc network, most of the nodes are mobile. This is the case, for instance, of the rescuers working in a disaster scenario as described above. – Relatively dispersed network: The adoption of the ad hoc networking paradigm is justified when the nodes composing the network are geographically dispersed. In fact, if network nodes are very close to each other, 1-hop wireless communication is usually possible and no multihop communication between nodes is necessary. Potential application of wireless ad hoc networks are numerous. Among them, we cite the following: – Fast traffic info delivery on highways and urban areas: Highways and urban areas can be equipped with fixed radio transmitters, which broadcast traffic information to cars equipped with GPS receivers passing close to a transmitter. In turn, the cars themselves act as relay of information so that the traffic updates can quickly reach AD HOC AND SENSOR NETWORKS 5 faraway drivers. As compared to traditional radio traffic info delivery, this technology will provide a much more accurate (localized) and faster service. – Ubiquitous Internet access: In a very near future (in part, this is already a reality), public areas such as airports, stations, shopping malls, and so on, will be equipped with wireless Internet access points. By using the portable devices of other users as wireless bridges, Internet access can be extended to virtually the entire urban area. – Delivery of location-aware information: By using fixed radio transmitters (for instance, the same transmitters used to broadcast traffic updates), location-aware informa- tion can be delivered to the interested users. Examples of location-aware informa- tion are tourist information, shows and events in the surrounding, information on shops/restaurants in the area, and so on. 1.1.2 Wireless sensor networks Wireless sensor networks (WSNs for short) are a particular type of ad hoc network, in which the nodes are ‘smart sensors’, that is, small devices (approximately the size of a coin) equipped with advanced sensing functionalities (thermal, pressure, acoustic, and so on, are examples of such sensing abilities), a small processor, and a short-range wireless transceiver. In this type of network, the sensors exchange information on the environment in order to build a global view of the monitored region, which is made accessible to the external user through one or more gateway node(s). Sensor networks are expected to bring a breakthrough in the way natural phenomena are observed: the accuracy of the observation will be considerably improved, leading to a better understanding and forecasting of such phenomena. The expected benefits to the community will be considerable. As in the case of ad hoc networks, to give a better idea of the potential of WSN technology, we describe in detail a sample application scenario. Consider a situation in which a WSN is used to monitor a vast and remote geographical region, in such a way abnormal events (e.g. a forest fire) can be quickly detected. In this scenario, smart sensors, each equipped with a battery, and significant processing and wireless communication capabilities, are placed in strategic positions, for example, on the top of a hill or in locations with wide view. Each sensor covers a few hectares area and can communicate with sensors in the surrounding. The sensor node gathers atmospheric data (temperature, pressure, humidity, wind velocity and direction) and analyzes atmosphere makeup to detect particular particles (e.g. ash). Furthermore, each sensor node is equipped with an infrared camera, which is able to detect thermal variations. Every sensor knows its geographic position, expressed in terms of degree of latitude and longitude. This can be accomplished either by equipping every node with a GPS receiver, or, since in this scenario sensor position is fixed, by setting the position in a sensor register at the time of deployment. Periodically, sensors exchange data with neighboring nodes in order to detect unusual situations that could be caused, for instance, by a starting fire (e.g. temperature at a sensor much higher than those of the neighbors). These ‘routine’ data are aggregated and propagated throughout the network and can be gathered by the external operator to collect atmospheric data (e.g. to check the air quality). When a potentially dangerous situation is detected (for instance, the infrared camera detects a rapid thermal increase in a certain zone), an emergency procedure is started: the 6 AD HOC AND SENSOR NETWORKS sensor node that has detected the abnormal condition communicates with its neighbors in order to verify whether the same condition has been detected by other sensors; then, it tries to accurately determine the geographic position of the hazard (if the same abnormal situation has been detected by other sensors, this can be accomplished using triangulation techniques; furthermore, the information on the wind velocity and direction can be useful both in the localization of the fire and in forecasting the direction of its propagation); once the position of the fire has been determined, an alarm message containing the fire’s geographic coordinates and (possibly) its propagation direction is disseminated with the maximum priority. This way, the external operator (for instance, a park ranger equipped with a portable device) is promptly alerted of the presence of fire, of its position, and of the forecasted propagation direction of the fire, and can intervene quickly. The fire-detection application scenario is summarized in Figure 1.1. We remark that this scenario has several interesting features, such as reduced impact on the environment (since sensor nodes have wireless transceivers, no wiring is needed), accuracy of coverage, and prompt alerting of the human operator. The above-described example outlines the features of a typical WSN application scenario: – Homogeneous network : Differing from the case of ad hoc networks, a WSN is typ- ically composed of nodes with the same features, especially for what concerns the communication apparatus. A partial exception to this rule is when different types of smart sensor nodes are used in the same network: for instance, a few ‘super nodes’ (with more memory and/or with a longer transmitting range) could be used in combina- tion with standard sensor nodes to increase the network monitoring ability. However, also in this case the number of different device classes used in the network is very limited (2–3 at most). Figure 1.1 Sensor network used for prompt fire detection. When a fire is detected, an alarm message (arrow) is generated by the sensor node(s) that detected the fire. The message is then propagated in the network until it reaches a park ranger. AD HOC AND SENSOR NETWORKS 7 Table 1.1 Comparison of typical features of wireless ad hoc and sensor networks Ad hoc Networks WSNs Heterogeneous devices Homogeneous devices Mobile nodes Stationary nodes Dispersed network Dispersed network Large network size – Stationary or quasistationary network : Differing from the case of ad hoc networks, nodes composing a WSN are typically stationary, or at most slowly moving. Given the very wide range of WSN applications, exceptions to this rule are possible. This is the case, for instance, of a sensor network used to track animal movements. – Relatively dispersed network : this feature is in common with ad hoc networks: a wireless sensor network is typically formed by nodes that are dispersed in a relatively large geographical region, so that 1-hop communication between nodes is, in general, not possible. – Large network size: Typically, the number of nodes composing a WSN is quite large, ranging from few tens to thousands of nodes. The differences/similarities between ad hoc and sensor networks are summarized in Table 1.1. Among the many possible WSN application scenarios, we cite the following: – Ocean temperature monitoring for improved weather forecast: It is known that the evolution of weather conditions is strongly influenced by the temperature of large water masses such as the oceans. However, nowadays our ability to perform a large- scale monitoring of the ocean temperature is scarce. Sensor networks can be used for this purpose. By dropping a large number of tiny sensors into the sea, water temperature and ocean currents can be accurately monitored, helping the scientists in the task of providing more accurate weather forecast. – Intrusion detection: Camera-equipped sensors can be used to form a network that monitors an area with restricted access. If the network is properly deployed, intruders can be detected and an alarm message quickly propagated to the external observer. – Avalanche prediction: Sensors equipped with location devices (such as GPS) can be used to monitor the movements of large snow masses, thus allowing a more accurate avalanche prediction. 1.2 Challenges Although the technology for ad hoc and sensor networks is relatively mature, the applications are almost completely lacking. This is in part due to the fact that some of the problems 8 AD HOC AND SENSOR NETWORKS related to ad hoc/sensor networking are still unsolved. In this section, we describe the state of progress of the current ad hoc and sensor network technology, and the main challenges that face the ad hoc/sensor network designer. 1.2.1 Ad hoc networks Wireless ad hoc networks have attracted the attention of researchers in academia and industry in the last few years. As a result of this considerable research activity, the basic mechanisms that enable wireless ad hoc communication have been designed and standardized. Just to cite the most popular examples, IEEE 802.11 (IEEE 1999) and Bluetooth (Bluetooth 1999) are communication standards that are implemented in a variety of commercial wireless equipment, and that allows infrastructure-less wireless communication between mutually compliant devices. Thus, wireless, multihop communication between different types of devices such as cell phones, laptops, PDAs, smart appliances, and so on, is possible with currently available technology. Despite the fact that the technology for ad hoc network exists, applications based on the ad hoc networking paradigm are almost completely lacking. This is because many of the challenges to be faced for a practical implementation of ad hoc network services are still to be solved. The main such challenges are the following: – Energy conservation: Since units in ad hoc networks are typically battery equipped, one of the primary design goals is to use this limited amount of energy as efficiently as possible. – Unstructured and/or time-varying network topology: Since the network nodes can, in principle, be arbitrarily placed in a certain region and are typically mobile, the topology of the graph that represents the wireless communication links between the nodes is usually unstructured. Furthermore, the network topology may vary with time, because of node mobility and/or failure. In these conditions, optimizing the performance of ad hoc network protocols is a very difficult task. – Low-quality communications: Communication on a wireless channel is, in general, much less reliable than in a wired channel. Furthermore, the quality of communica- tion is influenced by environmental factors (weather conditions, presence of obstacles, interference with other radio networks, etc.), which are time varying. Thus, applica- tions for ad hoc networks should be resilient to dramatically varying link conditions, tolerating also nonnegligible off-service time intervals of the wireless link. – Resource-constrained computation: Ad hoc networks are characterized by scarce resource availability; in particular, energy and network bandwidth are available in very limited amounts as compared to more traditional network paradigms. Protocols for ad hoc networks must strive to provide the desired performance level in spite of the few available resources. – Scalability: In some ad hoc network scenarios, the network can be composed of hundreds or thousands of nodes. This means that protocols for ad hoc networking must be able to operate efficiently in the presence of a very large number of nodes also. AD HOC AND SENSOR NETWORKS 9 In case of ad hoc networks used for ‘ubiquitous’ networking, the following issues must also be addressed: – Interoperability: In the ‘ubiquitous’ networking scenario described in Section 1.1.1, data should travel through the most diverse type of networks: ad hoc, cellular, satellite, wireless LAN, PSTN, Internet, and so on. Ideally, the user should smoothly switch from one network to the other without interrupting her applications. Implementing this sort of ‘network handoff’ is a very challenging task. – Definition of a feasible business model : Currently, accounting in wireless networks (cellular, and commercial wireless Internet access) is done at the base station, that is, using a centralized infrastructure. Furthermore, roaming is allowed only within networks of the same type (e.g. cell phone roaming when the user is in a foreign country). In the ‘ubiquitous’ scenario, it is still not clear which infrastructure should perform billing and which rules should be used to regulate roaming between different types of networks. – Stimulate cooperation between nodes: When designing a certain network protocol, it is usually assumed that all the nodes in the network voluntarily participate in the protocol execution. In some ad hoc network application scenarios, network nodes are owned by different authorities (private users, professionals, profit and/or nonprofit organizations, and so on), and voluntary participation in the protocol execution cannot be taken for granted. Thus, network nodes must be somehow stimulated to behave according to the protocol specifications. The issue of stimulating cooperation between nodes is treated in some detail in Chapter 16. 1.2.2 Wireless sensor networks In a manner similar to ad hoc networks, WSNs also have attracted the attention of both the academic and the industrial research community in the last few years. Firstly, a number of smart sensor prototypes have been designed and implemented by the academic research community. The most famous of such prototypes are probably the Berkeley Motes (Polastre et al. 2004) and Smart Dust (Pister 2001). Later on, many academic interdisciplinary projects have been funded (and are currently being funded) to actually deploy and utilize sensor networks. One such example is the Great Duck Island project, in which a WSN has been deployed to monitor the habitat of the nesting petrels without any human interference with animals (Mainwaring et al. 2002). Smart sensor nodes are also being produced and commercialized by some electronic manufacturer. We cite Crossbow, a company that produces on a large scale the Motes sensor nodes developed at UC Berkeley. Other major silicon companies such as Intel, Philips, Siemens, STMicrolectronics, and so on, are interested in the WSN technology, and are developing their own smart sensor node platform. There is also a considerable standardization activity in the field of WSNs. The most notable effort in this direction is the IEEE 802.15.4 standard currently under development, which defines the physical and MAC layer protocols for remote monitoring and control, as well as sensor network applications. ZigBee (ZigBeeAlliance 2004) is an industry con- sortium (currently involving more than 100 members, representing 22 countries on four continents) with the goal of promoting the IEEE 802.15.4 standard. [...]... and a company would have little or no profit in developing an application for a very specific scenario since the potential buyers would be very few 2 Modeling Ad Hoc Networks In this chapter, we introduce a simple but widely accepted model of ad hoc network Since sensor networks are a subclass of ad hoc networks, this model applies to this type of networks also 2. 1 The Wireless Channel Nodes in ad hoc. .. relative values In our simplified model, we assume that the radio is consuming conventional power 1 when the radio is idle, power 1.x when the radio is receiving a MODELING AD HOC NETWORKS 21 Table 2. 2 Nominal power consumption and transmit range of the CISCO IEEE 8 02. 11 a/b/g wireless card Power consumption is measured by the drain current, expressed in mA In the table, the minimum value of the nominal range... known Modeling path loss has historically been one of the most difficult tasks of the wireless system designer The mechanisms that regulate radio signal propagation in the environment can be grouped into three categories: reflection, diffraction and scattering Reflection occurs Topology Control in Wireless Ad Hoc and Sensor Networks P Santi  20 05 John Wiley & Sons, Ltd 14 MODELING AD HOC NETWORKS when... area is more and more irregular 2. 3 Modeling Energy Consumption One of the primary concerns of the ad hoc/ sensor networks designer is the efficient use of energy Thus, it is fundamental to model the node energy consumption accurately Since 20 MODELING AD HOC NETWORKS the features of typical nodes in ad hoc and sensor networks are quite different, we discuss energy models for the two classes of networks. ..10 AD HOC AND SENSOR NETWORKS Currently, we are in a phase in which the technology for implementing wireless sensor networks is relatively mature but applications based on sensor networks have not been completely defined In particular, industries strive to find significant markets for WSN applications The most promising ones seem to be home control, building automation, industrial automation, and automotive... Mbps), and the maximum value to the 6 Mbps data rate Power Idle (mA) 8 02. 11 a 8 02. 11 b/g Power Rx (mA) 20 3 20 3 20 3 554 539 530 318 327 28 2 Tx Range Indoor (m) 8 02. 11 a 8 02. 11 b 8 02. 11 g Power Tx (mA) Tx Range Outdoor (m) 13–50 27 –91 30–300 76–396 message, power 1.y when the radio is transmitting a message at full power, and power 0.z when the radio is in sleep mode (the actual values of x, y, and z... simulation/analysis: Since mobility models are used in the simulation of ad hoc networks, the model should be simple enough to be integrated in the MODELING AD HOC NETWORKS 23 simulator and to keep the simulation running time reasonable Furthermore, using relatively simple mobility models eases the task of deriving meaningful analytical results concerning fundamental network parameters in presence of mobility In turn,... implemented in network simulations tools such as Ns2 (Ns2 20 02) and GloMoSim (Zeng et al 1998) The Random Waypoint (RWP) model has been introduced in (Johnson and Maltz 1996) to study the performance of the DSR routing protocol In this model, each node chooses uniformly at random a destination point (the ‘waypoint’) within the deployment region R, and moves toward it along a straight line Node velocity... relatively simpler, as compared to the case of ad hoc networks In fact, sensor networks are 22 MODELING AD HOC NETWORKS Table 2. 3 Measured power consumption of a Rockwell’s WINS sensor node MCU Mode On On On On On On Sleep Sensor Mode Radio Mode Total Power (mW) On On On On On On On Tx (power 36.3 mW) Tx (power 0. 12 mW) Rx Idle Sleep Removed Removed 1080.5 771.1 751.6 727 .5 416.3 383.3 64.0 typically composed... used in very different situations, such as indoor or urban scenarios (ad hoc networks) , or under MODELING AD HOC NETWORKS 19 Figure 2. 3 Example of two-dimensional point graph Note that two of the links in the graph are unidirectional harsh conditions (sensor networks) In other words, in real-life situations, it is quite likely that the radio coverage region is highly irregular, because of the in uence . has Topology Control in Wireless Ad Hoc and Sensor Networks P. Santi  20 05 John Wiley & Sons, Ltd 4 AD HOC AND SENSOR NETWORKS devastated the city of Futuria destroying, among other things,. and the main challenges that face the ad hoc/ sensor network designer. 1 .2. 1 Ad hoc networks Wireless ad hoc networks have attracted the attention of researchers in academia and industry in the last. reflection, diffraction and scattering. Reflection occurs Topology Control in Wireless Ad Hoc and Sensor Networks P. Santi  20 05 John Wiley & Sons, Ltd 14 MODELING AD HOC NETWORKS when the electromagnetic