Chapter 5. AD HOC WIRELESS NETWORKS
5.1 INTRODUCTION
The principle behind ad hoc networking is multi-hop relaying, which traces its roots back to 500 B.C. Darius I (522-486 B.C.), the king of Persia, devised an innovative communication system that was used to send messages and news from his capital to the remote provinces of his empire by means of a line of shouting men positioned on tall structures or heights. This system was more than 25 times faster than normal messengers available at that time. The use of ad hoc voice communication was used in many ancient/tribal societies with a string of repeaters of drums, trumpets, or horns. In 1970, Norman Abramson and his fellow researchers at the University of Hawaii invented the ALOHAnet, an innovative communication system for linking together the universities of the Hawaiian islands. ALOHAnet utilized single-hop wireless packet switching and a multiple access solution for sharing a single channel. Even though ALOHAnet was originally implemented for a fixed single-hop wireless network, the basic idea was compelling and applicable to any environment where access to a common resource had to be negotiated among a set of uncoordinated nodes. The success and novelty of ALOHAnet triggered widespread interest in different directions of computer communication, including the work that led to the development of Ethernet by Robert Metcalfe and the packet radio network (PRNET) project sponsored by the defense advanced research projects agency (DARPA) [1]. The PRNET project was aimed at developing a packet wireless network for military applications. Even though the initial attempt had a centralized control, it quickly evolved into a distributed multi-hop wireless communication system that could operate over a large geographical area. Each mobile node had a broadcast radio interface that provided many advantages such as the use of a single channel, simpler channel management techniques, and the ease of supporting mobility. PRNET used a combination of ALOHA and carrier sense multiple access (CSMA) for access to the shared radio channel. The radio interface employed the direct-sequence (DS) spread spectrum scheme. The system was designed to selforganize, self-configure, and detect radio connectivity for the dynamic operation of a routing protocol without any support from fixed infrastructure. The major issues that the PRNET project faced include those of obtaining, maintaining, and utilizing the topology information, error and flow control over the wireless links, reconfiguration of paths to handle path breaks arising due to the mobility of nodes and routers, processing and storage capability of nodes, and distributed channel sharing. The successful demonstrations of the PRNET proved the feasibility and efficiency of infrastructure-less networks and their applications for civilian and military purposes. DARPA extended the work on multi-hop wireless networks through the survivable radio networks (SURAN) project that aimed at providing ad hoc networking with small, low-cost, low-power devices with efficient protocols and improved scalability and survivability (the ability of a network to survive the failure of network nodes and links). During the 1980s, research on military applications was extensively funded across the globe. Realizing the necessity of open standards in this emerging area of computer communication, a working group within the Internet Engineering Task Force (IETF), termed the mobile ad hoc networks (MANET) working group [2], was formed to standardize the protocols and functional specifications of ad hoc wireless networks. The vision of the IETF effort in the MANET working group is to provide improved standardized routing functionality to support self-organizing mobile networking infrastructure.
In 1994, the Swedish communication equipment maker Ericsson proposed to develop a short-range, low-power, low-complexity, and inexpensive radio interface and associated communication protocols referred to as Bluetooth for ubiquitous connectivity among heterogeneous devices, as discussed in Section 2.5. This effort was later taken over by a Special Interest Group (SIG) formed by several major computer and telecommunication vendors such as 3Com, Ericsson, IBM, Intel, Lucent, Microsoft, Motorola, Nokia, and Toshiba. The Bluetooth SIG aims at delivering a universal solution for connectivity among heterogeneous devices. This is one of the first commercial realizations of ad hoc wireless networking. Bluetooth standardizes the single-hop point-to-point wireless link that helps in exchanging voice or data, and formation of piconets that are formed by a group of nodes in a smaller geographical region where every node can reach every other node in the group within a single-hop. Multiple piconets can form a scatternet, which necessitates the use of multi-hop routing protocols.
Even though ad hoc wireless networks are expected to work in the absence of any fixed infrastructure, recent advances in wireless network architectures reveal interesting solutions that enable the mobile ad hoc nodes to function in the presence of infrastructure. Multi-hop cellular networks (MCNs) [3] and self-organizing packet radio ad hoc networks with overlay (SOPRANO) [4] are examples of such types of networks. These hybrid architectures (which combine the benefits of cellular and ad hoc wireless networks) improve the capacity of the system significantly. Even with all the promises that are offered by ad hoc wireless networks, successful commercial deployment requires realistic solutions to different problems, including support for QoS provisioning and real-time applications, pricing, cooperative functioning, energy-efficient relaying, load balancing, and support for multicast traffic.
5.1.1 Cellular and Ad Hoc Wireless Networks
Figure 5.1 shows a representation of different wireless networks. The current cellular wireless networks (depicted in Figure 5.2) are classified as the infrastructure dependent networks. The path setup for a call between two nodes, say, node C to node E, is completed through the base station as illustrated in Figure 5.2.
Figure 5.1. Cellular and ad hoc wireless networks.
Figure 5.2. A cellular network.
Ad hoc wireless networks are defined as the category of wireless networks that utilize multi-hop radio relaying and are capable of operating without the support of any fixed infrastructure (hence they are also called infrastructureless networks). The absence of any central coordinator or base station makes the routing a complex one compared to cellular networks. Ad hoc wireless network topology for the cellular network shown in Figure 5.2 is illustrated in Figure 5.3. Note that in Figure 5.3 the cell boundaries are shown purely for comparison with the cellular network in Figure 5.2 and do not carry any special significance. The path setup for a call between two nodes, say, node C to node E, is completed through the intermediate mobile node F, as illustrated in Figure 5.3. Wireless mesh networks and wireless sensor networks are specific examples of ad hoc wireless networks.
Figure 5.3. An ad hoc wireless network.
The major differences between cellular networks and ad hoc wireless networks are summarized in Table 5.1. The presence of base stations simplifies routing and resource management in a cellular network as the routing decisions are made in a centralized manner with more information about the destination node. But in an ad hoc wireless network, the routing and resource management are done in a distributed manner in which all nodes coordinate to enable communication among themselves. This requires each node to be more intelligent so that it can function both as a network host for transmitting and receiving data and as a network router for routing packets from other nodes. Hence the mobile nodes in ad hoc wireless networks are more complex than their counterparts in cellular networks.
Table 5.1. Differences between cellular networks and ad hoc wireless networks
5.1.2 Applications of Ad Hoc Wireless Networks
Ad hoc wireless networks, due to their quick and economically less demanding deployment, find applications in several areas. Some of these include: military applications, collaborative and distributed computing, emergency operations, wireless mesh networks, wireless sensor networks, and hybrid wireless network architectures.
Military Applications
Ad hoc wireless networks can be very useful in establishing communication among a group of soldiers for tactical operations. Setting up a fixed infrastructure for communication among a group of soldiers in enemy territories or in inhospitable terrains may not be possible. In such environments, ad hoc wireless networks provide the required communication mechanism quickly. Another application in this area can be the coordination of military objects moving at high speeds such as fleets of airplanes or warships. Such applications require quick and reliable communication. Secure communication is of prime importance as eavesdropping or other security threats can compromise the purpose of communication or the safety of personnel involved in these tactical operations. They also require the support of reliable and secure multimedia multicasting. For example, the leader of a group of soldiers may want to give an order to all the soldiers or to a set of selected personnel involved in the operation. Hence, the routing protocol in these applications should be able to provide quick, secure, and reliable multicast communication with support for real-time traffic.
As the military applications require very secure communication at any cost, the vehicle-mounted nodes can be assumed to be very sophisticated and powerful. They can have multiple high-power transceivers, each with the ability to hop between different frequencies for security reasons. Such communication systems can be assumed to be equipped with long-life batteries that might not be economically viable for normal usage. They can even use other services such as location tracking [using the global positioning system (GPS)] or other satellite-based services for efficient communication and coordination. Resource constraints such as battery life and transmitting power may not exist in certain types of applications of ad hoc wireless networks. For example, the ad hoc wireless network formed by a fleet of military tanks may not suffer from the power source constraints present in the ad hoc network formed by a set of wearable devices used by the foot soldiers.
In short, the primary nature of the communication required in a military environment enforces certain important requirements on ad hoc wireless networks, namely, reliability, efficiency, secure communication, and support for multicast routing.
Collaborative and Distributed Computing
Another domain in which the ad hoc wireless networks find applications is collaborative computing. The requirement of a temporary communication infrastructure for quick communication with minimal configuration among a group of people in a conference or gathering necessitates the formation of an ad hoc wireless network. For example, consider a group of researchers who want to share their research findings or presentation materials during a conference, or a lecturer distributing notes to the class on the fly. In such cases, the formation of an ad hoc wireless network with the necessary support for reliable multicast routing can serve the purpose. The distributed file sharing applications utilized in such situations do not require the level of security expected in a military environment. But the reliability of data transfer is of high importance. Consider the example where a node that is part of an ad hoc wireless network has to distribute a file to other nodes in the network. Though this application does not demand the communication to be interruption-free, the goal of the transmission is that all the desired receivers must have the replica of the transmitted file. Other applications such as streaming of multimedia objects among the participating nodes in an ad hoc wireless network may require support for soft real-time communication. The users of such applications prefer economical and portable devices, usually powered by battery sources. Hence, a mobile node may drain its battery and can have varying transmission power, which may result in unidirectional links with its neighbors. Devices used for such applications could typically be laptops with add-on wireless interface cards, enhanced personal digital assistants (PDAs), or mobile devices with high processing power. In the presence of such heterogeneity, interoperability is an important issue.
Emergency Operations
Ad hoc wireless networks are very useful in emergency operations such as search and rescue, crowd control, and commando operations. The major factors that favor ad hoc wireless networks for such tasks are self-configuration of the system with minimal overhead, independent of fixed or centralized infrastructure, the nature of the terrain of such applications, the freedom and flexibility of mobility, and the unavailability of conventional communication infrastructure. In environments where the conventional infrastructure-based communication facilities are destroyed due to a war or due to natural calamities such as earthquakes, immediate deployment of ad hoc wireless networks would be a good solution for coordinating rescue activities. Since the ad hoc wireless networks require minimum initial network configuration for their functioning, very little or no delay is involved in making the network fully operational. The above-mentioned scenarios are unexpected, in most cases unavoidable, and can affect a large number of people. Ad hoc wireless networks employed in such circumstances should be distributed and scalable to a large number of nodes. They should also be able to provide fault-tolerant communication paths. Real-time communication capability is also important since voice communication predominates data communication in such situations.
Wireless Mesh Networks
Wireless mesh networks are ad hoc wireless networks that are formed to provide an alternate communication infrastructure for mobile or fixed nodes/users, without the spectrum reuse constraints and the requirements of network planning of cellular networks. The mesh topology of wireless mesh networks provides many alternate paths for a data transfer session between a source and destination, resulting in quick reconfiguration of the path when the existing path fails due to node failures. Wireless mesh networks provide the most economical data transfer capability coupled with the freedom of mobility. Since the infrastructure built is in the form of small radio relaying devices fixed on the rooftops of the houses in a residential zone as shown in Figure 5.4, or similar devices fitted on the lamp posts as depicted in Figure 5.5, the investment required in wireless mesh networks is much less than what is required for the cellular network counterparts. Such networks are formed by placing wireless relaying equipment spread across the area to be covered by the network. The possible deployment scenarios of wireless mesh networks include: residential zones (where broadband Internet connectivity is required), highways (where a communication facility for moving automobiles is required), business zones (where an alternate communication system to cellular networks is required), important civilian regions (where a high degree of service availability is required), and university campuses (where inexpensive campus-wide network coverage can be provided). Wireless mesh networks should be capable of self-organization and maintenance. The ability of the network to overcome single or multiple node failures resulting from disasters makes it convenient for providing the communication infrastructure for strategic applications. The major advantages of wireless mesh networks are support for a high data rate, quick and low cost of deployment, enhanced services, high scalability, easy extendability, high availability, and low cost per bit. Wireless mesh networks operate at the license-free ISM bands around 2.4 GHz and 5 GHz. Depending on the technology used for the physical layer and MAC layer communication, data rates of 2 Mbps to 60 Mbps can be supported. For example, if IEEE 802.11a is used, a maximum data rate of 54 Mbps can be supported. The deployment time required for this network is much less than that provided by other infrastructure-based networks. Incremental deployment or partial batch deployment can also be done. Wireless mesh networks provide a very economical communication infrastructure in terms of both deployment and data transfer costs. Services such as smart environments that update information about the environment or locality to the visiting nodes are also possible in such an environment. A truck driver can utilize enhanced location discovery services, and hence spotting his location on an updated digital map is possible. Mesh networks scale well to provide support to a large number of nodes. Even at a very high density of mobile nodes, by employing power control at the mobile nodes and relay nodes, better system throughput and support for a large number of users can be achieved. But in the case of cellular networks, improving scalability requires additional infrastructural nodes, which in turn involves high cost. As mentioned earlier, mesh networks provide expandability of service in a cost-effective manner. Partial roll out and commissioning of the network and extending the service in a seamless manner without affecting the existing installation are the benefits from the viewpoint of service providers. Wireless mesh networks provide very high availability compared to the existing cellular architecture, where the presence of a fixed base station that covers a much larger area involves the risk of a single point of failure.
Figure 5.4. Wireless mesh network operating in a residential zone.
Figure 5.5. Wireless mesh network covering a highway.
Wireless Sensor Networks
Sensor networks are a special category of ad hoc wireless networks that are used to provide a wireless communication infrastructure among the sensors deployed in a specific application domain. Recent advances in wireless communication technology and research in ad hoc wireless networks have made smart sensing a reality. Sensor nodes are tiny devices that have the capability of sensing physical parameters, processing the data gathered, and communicating over the network to the monitoring station. A sensor network is a collection of a large number of sensor nodes that are deployed in a particular region. The activity of sensing can be periodic or sporadic. An example for the periodic type is the sensing of environmental factors for the measurement of parameters such as temperature, humidity, and nuclear radiation. Detecting border intrusion, sensing the temperature of a furnace to prevent it rising beyond a threshold, and measuring the stress on critical structures or machinery are examples of the sensing activities that belong to the sporadic type. Some of the domains of application for sensor networks are military, health care, home security, and environmental monitoring. The issues that make sensor networks a distinct category of ad hoc wireless networks are the following:
• Mobility of nodes: Mobility of nodes is not a mandatory requirement in sensor networks. For example, the nodes deployed for periodic monitoring of soil properties are not required to be mobile. However, the sensor nodes that are fitted on the bodies of patients in a post-surgery ward of a hospital may be designed to support limited or partial mobility. In general, sensor networks need not in all cases be designed to support mobility of sensor nodes.
• Size of the network: The number of nodes in the sensor network can be much larger than that in a typical ad hoc wireless network.
• Density of deployment: The density of nodes in a sensor network varies with the domain of application. For example, military applications require high availability of the network, making redundancy a high priority.
• Power constraints: The power constraints in sensor networks are much more stringent than those in ad hoc wireless networks. This is mainly because the sensor nodes are expected to operate in harsh environmental or geographical conditions, with minimum or no human supervision and maintenance. In certain cases, the recharging of the energy source is impossible. Running such a network, with nodes powered by a battery source with limited energy, demands very efficient protocols at network, data link, and physical layer. The power sources used in sensor networks can be classified into the following three categories:
– Replenishable power source: In certain applications of sensor networks, the power source can be replaced when the existing source is fully drained (e.g., wearable sensors that are used to sense body parameters).
– Non-replenishable power source: In some specific applications of sensor networks, the power source cannot be replenished once the network has been deployed. The replacement of the sensor node is the only solution to it (e.g., deployment of sensor nodes in a remote, hazardous terrain).
– Regenerative power source: Power sources employed in sensor networks that belong to this category have the capability of regenerating power from the physical parameter under measurement. For example, the sensor employed for sensing temperature at a power plant can use power sources that can generate power by using appropriate transducers.
• Data/information fusion: The limited bandwidth and power constraints demand aggregation of bits and information at the intermediate relay nodes that are responsible for relaying. Data fusion refers to the aggregation of multiple packets into one before relaying it. This mainly aims at reducing the bandwidth consumed by redundant headers of the packets and reducing the media access delay involved in transmitting multiple packets. Information fusion aims at processing the sensed data at the intermediate nodes and relaying the outcome to the monitor node.
• Traffic distribution: The communication traffic pattern varies with the domain of application in sensor networks. For example, the environmental sensing application generates short periodic packets indicating the status of the environmental parameter under observation to a central monitoring station. This kind of traffic demands low bandwidth. The sensor network employed in detecting border intrusions in a military application generates traffic on detection of certain events; in most cases these events might have time constraints for delivery. In contrast, ad hoc wireless networks generally carry user traffic such as digitized and packetized voice stream or data traffic, which demands higher bandwidth.
Hybrid Wireless Networks
One of the major application areas of ad hoc wireless networks is in hybrid wireless architectures such as multi-hop cellular networks (MCNs) [5] and [3] and integrated cellular ad hoc relay (iCAR) networks [6]. The tremendous growth in the subscriber base of existing cellular networks has shrunk the cell size up to the pico-cell level. The primary concept behind cellular networks is geographical channel reuse. Several techniques such as cell sectoring, cell resizing, and multi-tier cells have been proposed to increase the capacity of cellular networks. Most of these schemes also increase the equipment cost. The capacity (maximum throughput) of a cellular network can be increased if the network incorporates the properties of multi-hop relaying along with the support of existing fixed infrastructure. MCNs combine the reliability and support of fixed base stations of cellular networks with flexibility and multi-hop relaying of ad hoc wireless networks.
The MCN architecture is depicted in Figure 5.6. In this architecture, when two nodes (which are not in direct transmission range) in the same cell want to communicate with each other, the connection is routed through multiple wireless hops over the intermediate nodes. The base station maintains the information about the topology of the network for efficient routing. The base station may or may not be involved in this multi-hop path. Suppose node A wants to communicate with node B. If all nodes are capable of operating in MCN mode, node A can reach node B directly if the node B is within node A's transmission range. When node C wants to communicate with node E and both are in the same cell, node C can reach node E through node D, which acts as an intermediate relay node. Such hybrid wireless networks can provide high capacity resulting in lowering the cost of communication to less than that in single-hop cellular networks. The major advantages of hybrid wireless networks are as follows:
• Higher capacity than cellular networks obtained due to the better channel reuse provided by reduction of transmission power, as mobile nodes use a power range that is a fraction of the cell radius.
• Increased flexibility and reliability in routing. The flexibility is in terms of selecting the best suitable nodes for routing, which is done through multiple mobile nodes or through base stations, or by a combination of both. The increased reliability is in terms of resilience to failure of base stations, in which case a node can reach other nearby base stations using multi-hop paths.
• Better coverage and connectivity in holes (areas that are not covered due to transmission difficulties such as antenna coverage or the direction of antenna) of a cell can be provided by means of multiple hops through intermediate nodes in the cell.
Figure 5.6. MCN architecture.
5.2 ISSUES IN AD HOC WIRELESS NETWORKS
This section discusses the major issues and challenges that need to be considered when an ad hoc wireless system is to be designed. The deployment considerations for installation, operation, and maintenance of ad hoc wireless networks are also provided. The major issues that affect the design, deployment, and performance of an ad hoc wireless system are as follows:
• Medium access scheme
• Routing
• Multicasting
• Transport layer protocol
• Pricing scheme
• Quality of service provisioning
• Self-organization
• Security
• Energy management
• Addressing and service discovery
• Scalability
• Deployment considerations
5.2.1 Medium Access Scheme
The primary responsibility of a medium access control (MAC) protocol in ad hoc wireless networks is the distributed arbitration for the shared channel for transmission of packets. The performance of any wireless network hinges on the MAC protocol, more so for ad hoc wireless networks. The major issues to be considered in designing a MAC protocol for ad hoc wireless networks are as follows:
• Distributed operation: The ad hoc wireless networks need to operate in environments where no centralized coordination is possible. The MAC protocol design should be fully distributed involving minimum control overhead. In the case of polling-based MAC protocols, partial coordination is required.
• Synchronization: The MAC protocol design should take into account the requirement of time synchronization. Synchronization is mandatory for TDMA-based systems for management of transmission and reception slots. Synchronization involves usage of scarce resources such as bandwidth and battery power. The control packets used for synchronization can also increase collisions in the network.
• Hidden terminals: Hidden terminals are nodes that are hidden (or not reachable) from the sender of a data transmission session, but are reachable to the receiver of the session. In such cases, the hidden terminal can cause collisions at the receiver node. The presence of hidden terminals can significantly reduce the throughput of a MAC protocol used in ad hoc wireless networks. Hence the MAC protocol should be able to alleviate the effects of hidden terminals.
• Exposed terminals: Exposed terminals, the nodes that are in the transmission range of the sender of an on-going session, are prevented from making a transmission. In order to improve the efficiency of the MAC protocol, the exposed nodes should be allowed to transmit in a controlled fashion without causing collision to the on-going data transfer.
• Throughput: The MAC protocol employed in ad hoc wireless networks should attempt to maximize the throughput of the system. The important considerations for throughput enhancement are minimizing the occurrence of collisions, maximizing channel utilization, and minimizing control overhead.
• Access delay: The access delay refers to the average delay that any packet experiences to get transmitted. The MAC protocol should attempt to minimize the delay.
• Fairness: Fairness refers to the ability of the MAC protocol to provide an equal share or weighted share of the bandwidth to all competing nodes. Fairness can be either node-based or flow-based. The former attempts to provide an equal bandwidth share for competing nodes whereas the latter provides an equal share for competing data transfer sessions. In ad hoc wireless networks, fairness is important due to the multi-hop relaying done by the nodes. An unfair relaying load for a node results in draining the resources of that node much faster than that of other nodes.
• Real-time traffic support: In a contention-based channel access environment, without any central coordination, with limited bandwidth, and with location-dependent contention, supporting time-sensitive traffic such as voice, video, and real-time data requires explicit support from the MAC protocol.
• Resource reservation: The provisioning of QoS defined by parameters such as bandwidth, delay, and jitter requires reservation of resources such as bandwidth, buffer space, and processing power. The inherent mobility of nodes in ad hoc wireless networks makes such reservation of resources a difficult task. A MAC protocol should be able to provide mechanisms for supporting resource reservation and QoS provisioning.
• Ability to measure resource availability: In order to handle the resources such as bandwidth efficiently and perform call admission control based on their availability, the MAC protocol should be able to provide an estimation of resource availability at every node. This can also be used for making congestion-control decisions.
• Capability for power control: The transmission power control reduces the energy consumption at the nodes, causes a decrease in interference at neighboring nodes, and increases frequency reuse. Support for power control at the MAC layer is very important in the ad hoc wireless environment.
• Adaptive rate control: This refers to the variation in the data bit rate achieved over a channel. A MAC protocol that has adaptive rate control can make use of a high data rate when the sender and receiver are nearby and adaptively reduce the data rate as they move away from each other.
• Use of directional antennas: This has many advantages that include increased spectrum reuse, reduction in interference, and reduced power consumption. Most of the existing MAC protocols that use omnidirectional antennas do not work with directional antennas.
5.2.2 Routing
The responsibilities of a routing protocol include exchanging the route information; finding a feasible path to a destination based on criteria such as hop length, minimum power required, and lifetime of the wireless link; gathering information about the path breaks; mending the broken paths expending minimum processing power and bandwidth; and utilizing minimum bandwidth. The major challenges that a routing protocol faces are as follows:
• Mobility: One of the most important properties of ad hoc wireless networks is the mobility associated with the nodes. The mobility of nodes results in frequent path breaks, packet collisions, transient loops, stale routing information, and difficulty in resource reservation. A good routing protocol should be able to efficiently solve all the above issues.
• Bandwidth constraint: Since the channel is shared by all nodes in the broadcast region (any region in which all nodes can hear all other nodes), the bandwidth available per wireless link depends on the number of nodes and the traffic they handle. Thus only a fraction of the total bandwidth is available for every node.
• Error-prone and shared channel: The bit error rate (BER) in a wireless channel is very high (of the order of 10-5 to 10-3) compared to that in its wired counterparts (of the order of 10-12 to 10-9). Routing protocols designed for ad hoc wireless networks should take this into account. Consideration of the state of the wireless link, signal-to-noise ratio, and path loss for routing in ad hoc wireless networks can improve the efficiency of the routing protocol.
• Location-dependent contention: The load on the wireless channel varies with the number of nodes present in a given geographical region. This makes the contention for the channel high when the number of nodes increases. The high contention for the channel results in a high number of collisions and a subsequent wastage of bandwidth. A good routing protocol should have built-in mechanisms for distributing the network load uniformly across the network so that the formation of regions where channel contention is high can be avoided.
• Other resource constraints: The constraints on resources such as computing power, battery power, and buffer storage also limit the capability of a routing protocol.
The major requirements of a routing protocol in ad hoc wireless networks are the following:
• Minimum route acquisition delay: The route acquisition delay for a node that does not have a route to a particular destination node should be as minimal as possible. This delay may vary with the size of the network and the network load.
• Quick route reconfiguration: The unpredictable changes in the topology of the network require that the routing protocol be able to quickly perform route reconfiguration in order to handle path breaks and subsequent packet losses.
• Loop-free routing: This is a fundamental requirement of any routing protocol to avoid unnecessary wastage of network bandwidth. In ad hoc wireless networks, due to the random movement of nodes, transient loops may form in the route thus established. A routing protocol should detect such transient routing loops and take corrective actions.
• Distributed routing approach: An ad hoc wireless network is a fully distributed wireless network and the use of centralized routing approaches in such a network may consume a large amount of bandwidth.
• Minimum control overhead: The control packets exchanged for finding a new route and maintaining existing routes should be kept as minimal as possible. The control packets consume precious bandwidth and can cause collisions with data packets, thereby reducing network throughput.
• Scalability: Scalability is the ability of the routing protocol to scale well (i.e., perform efficiently) in a network with a large number of nodes. This requires minimization of control overhead and adaptation of the routing protocol to the network size.
• Provisioning of QoS: The routing protocol should be able to provide a certain level of QoS as demanded by the nodes or the category of calls. The QoS parameters can be bandwidth, delay, jitter, packet delivery ratio, and throughput. Supporting differentiated classes of service may be of importance in tactical operations.
• Support for time-sensitive traffic: Tactical communications and similar applications require support for time-sensitive traffic. The routing protocol should be able to support both hard real-time and soft real-time traffic.
• Security and privacy: The routing protocol in ad hoc wireless networks must be resilient to threats and vulnerabilities. It must have inbuilt capability to avoid resource consumption, denial-of-service, impersonation, and similar attacks possible against an ad hoc wireless network.
5.2.3 Multicasting
Multicasting plays an important role in the typical applications of ad hoc wireless networks, namely, emergency search-and-rescue operations and military communication. In such an environment, nodes form groups to carry out certain tasks that require point-to-multipoint and multipoint-to-multipoint voice and data communication. The arbitrary movement of nodes changes the topology dynamically in an unpredictable manner. The mobility of nodes, with the constraints of power source and bandwidth, makes multicast routing very challenging. Traditional wired network multicast protocols such as core based trees (CBT), protocol independent multicast (PIM), and distance vector multicast routing protocol (DVMRP) do not perform well in ad hoc wireless networks because a tree-based multicast structure is highly unstable and needs to be frequently readjusted to include broken links. Use of any global routing structure such as the link-state table results in high control overhead. The use of single-link connectivity among the nodes in a multicast group results in a tree-shaped multicast routing topology. Such a tree-shaped topology provides high multicast efficiency, with low packet delivery ratio due to the frequent tree breaks. Provisioning of multiple links among the nodes in an ad hoc wireless network results in a mesh-shaped structure. The mesh-based multicast routing structure may work well in a high-mobility environment. The major issues in designing multicast routing protocols are as follows:
• Robustness: The multicast routing protocol must be able to recover and reconfigure quickly from potential mobility-induced link breaks thus making it suitable for use in highly dynamic environments.
• Efficiency: A multicast protocol should make a minimum number of transmissions to deliver a data packet to all the group members.
• Control overhead: The scarce bandwidth availability in ad hoc wireless networks demands minimal control overhead for the multicast session.
• Quality of service: QoS support is essential in multicast routing because, in most cases, the data transferred in a multicast session is time-sensitive.
• Efficient group management: Group management refers to the process of accepting multicast session members and maintaining the connectivity among them until the session expires. This process of group management needs to be performed with minimal exchange of control messages.
• Scalability: The multicast routing protocol should be able to scale for a network with a large number of nodes.
• Security: Authentication of session members and prevention of non-members from gaining unauthorized information play a major role in military communications.
5.2.4 Transport Layer Protocols
The main objectives of the transport layer protocols include setting up and maintaining end-to-end connections, reliable end-to-end delivery of data packets, flow control, and congestion control. There exist simple connectionless transport layer protocols (e.g., UDP) which neither perform flow control and congestion control nor provide reliable data transfer. Such unreliable connectionless transport layer protocols do not take into account the current network status such as congestion at the intermediate links, the rate of collision, or other similar factors affecting the network throughput. This behavior of the transport layer protocols increases the contention of the already-choked wireless links. For example, in an ad hoc wireless network that employs a contention-based MAC protocol, nodes in a high-contention region experience several backoff states, resulting in an increased number of collisions and a high latency. Connectionless transport layer protocols, unaware of this situation, increase the load in the network, degrading the network performance.
The major performance degradation faced by a reliable connection-oriented transport layer protocol such as transmission control protocol (TCP) in an ad hoc wireless network arises due to frequent path breaks, presence of stale routing information, high channel error rate, and frequent network partitions.
Further discussion of each of the above properties and their effect on the performance of the transport layer protocol assumes TCP as the transport layer protocol. Due to the mobility of nodes and limited transmission range, an existing path to a destination node experiences frequent path breaks. Each path break results in route reconfiguration that depends on the routing protocol employed. The process of finding an alternate path or reconfiguring the broken path might take longer than the retransmission timeout of the transport layer at the sender, resulting in retransmission of packets and execution of the congestion control algorithm. The congestion control algorithm decreases the size of the congestion window, resulting in low throughput. In an environment where path breaks are frequent, the execution of congestion control algorithms on every path break affects the throughput drastically.
The latency associated with the reconfiguration of a broken path and the use of route caches result in stale route information at the nodes. Hence the packets will be forwarded through multiple paths to a destination, causing an increase in the number of out-of-order packets. Also, multipath routing protocols such as temporally-ordered routing algorithm (TORA) [7] and split multipath routing (SMR) protocols ([8] and [9]) employ multiple paths between a source-destination pair. Out-of-order packet arrivals force the receiver of the TCP connection to generate duplicate acknowledgments (ACKs). On receiving duplicate ACKs, the sender invokes the congestion control algorithm.
Wireless channels are inherently unreliable due to the high probability of errors caused by interference. In addition to the error due to the channel noise, the presence of hidden terminals also contributes to the increased loss of TCP data packets or ACKs. When the TCP ACK is delayed more than the round-trip timeout, the congestion control algorithm is invoked.
Due to the mobility of the nodes, ad hoc wireless networks frequently experience isolation of nodes from the rest of the network or occurrence of partitions in the network. If a TCP connection spans across multiple partitions, that is, the sender and receiver of the connection are in two different partitions, all the packets get dropped. This tends to be more serious when the partitions exist for a long duration, resulting in multiple retransmissions of the TCP packets and subsequent increase in the retransmission timers. Such a behavior causes long periods of inactivity even when a transient partition in the network lasts for a short while.
Adaptation of the existing transport layer protocols should attempt to handle the above issues for performing efficiently in ad hoc wireless networks.
5.2.5 Pricing Scheme
An ad hoc wireless network's functioning depends on the presence of relaying nodes and their willingness to relay other nodes' traffic. Even if the node density is sufficient enough to ensure a fully connected network, a relaying neighbor node may not be interested in relaying a call and may just decide to power down. Assume that an optimal route from node A to node B passes through node C, and node C is not powered on. Then node A will have to set up a costlier and non-optimal route to B. The non-optimal path consumes more resources and affects the throughput of the system. As the intermediate nodes in a path that relay the data packets expend their resources such as battery charge and computing power, they should be properly compensated. Hence pricing schemes that incorporate service compensation or service reimbursement are required. Ad hoc wireless networks employed for special tasks such as military missions, rescue operations, and law enforcement do not require such pricing schemes, whereas the successful commercial deployment of ad hoc wireless networks requires billing and pricing. The obvious solution to provide participation guarantee is to provide incentives to forwarding nodes.
5.2.6 Quality of Service Provisioning
Quality of service (QoS) is the performance level of services offered by a service provider or a network to the user. QoS provisioning often requires negotiation between the host and the network, resource reservation schemes, priority scheduling, and call admission control. Rendering QoS in ad hoc wireless networks can be on a per flow, per link, or per node basis. In ad hoc wireless networks, the boundary between the service provider (network) and the user (host) is blurred, thus making it essential to have better coordination among the hosts to achieve QoS. The lack of central coordination and limited resources exacerbate the problem. In this section, a brief discussion of QoS parameters, QoS-aware routing, and QoS frameworks in ad hoc wireless networks is provided.
• QoS parameters: As different applications have different requirements, their level of QoS and the associated QoS parameters also differ from application to application. For example, for multimedia applications, the bandwidth and delay are the key parameters, whereas military applications have the additional requirements of security and reliability. For defense applications, finding trustworthy intermediate hosts and routing through them can be a QoS parameter. For applications such as emergency search-and-rescue operations, availability is the key QoS parameter. Multiple link disjoint paths can be the major requirement for such applications. Applications for hybrid wireless networks can have maximum available link life, delay, channel utilization, and bandwidth as the key parameters for QoS. Finally, applications such as communication among the nodes in a sensor network require that the transmission among them results in minimum energy consumption, hence battery life and energy conservation can be the prime QoS parameters here.
• QoS-aware routing: The first step toward a QoS-aware routing protocol is to have the routing use QoS parameters for finding a path. The parameters that can be considered for routing decisions are network throughput, packet delivery ratio, reliability, delay, delay jitter, packet loss rate, bit error rate, and path loss. Decisions on the level of QoS and the related parameters for such services in ad hoc wireless networks are application-specific and are to be met by the underlying network. For example, in the case where the QoS parameter is bandwidth, the routing protocol utilizes the available bandwidth at every link to select a path with necessary bandwidth. This also demands the capability to reserve the required amount of bandwidth for that particular connection.
• QoS framework: A framework for QoS is a complete system that attempts to provide the promised services to each user or application. All the components within this subsystem should cooperate in providing the required services. The key component of QoS framework is a QoS service model which defines the way user requirements are served. The key design issue is whether to serve the user on a per-session basis or a per-class basis. Each class represents an aggregation of users based on certain criteria. The other key components of this framework are QoS routing to find all or some feasible paths in the network that can satisfy user requirements, QoS signaling for resource reservation required by the user or application, QoS medium access control, connection admission control, and scheduling schemes pertaining to that service model. The QoS modules such as routing protocol, signaling protocol, and resource management should react promptly according to changes in the network state (topology change in ad hoc wireless networks) and flow state (change in end-to-end view of service delivered).
5.2.7 Self-Organization
One very important property that an ad hoc wireless network should exhibit is organizing and maintaining the network by itself. The major activities that an ad hoc wireless network is required to perform for self-organization are neighbor discovery, topology organization, and topology reorganization. During the neighbor discovery phase, every node in the network gathers information about its neighbors and maintains that information in appropriate data structures. This may require periodic transmission of short packets named beacons, or promiscuous snooping on the channel for detecting activities of neighbors. Certain MAC protocols permit varying the transmission power to improve upon spectrum reusability. In the topology organization phase, every node in the network gathers information about the entire network or a part of the network in order to maintain topological information.
During the topology reorganization phase, the ad hoc wireless networks require updating the topology information by incorporating the topological changes occurred in the network due to the mobility of nodes, failure of nodes, or complete depletion of power sources of the nodes. The reorganization consists of two major activities. First is the periodic or aperiodic exchange of topological information. Second is the adaptability (recovery from major topological changes in the network).
Similarly, network partitioning and merging of two existing partitions require major topological reorganization. Ad hoc wireless networks should be able to perform self-organization quickly and efficiently in a way transparent to the user and the application.
5.2.8 Security
The security of communication in ad hoc wireless networks is very important, especially in military applications. The lack of any central coordination and shared wireless medium makes them more vulnerable to attacks than wired networks. The attacks against ad hoc wireless networks are generally classified into two types: passive and active attacks. Passive attacks refer to the attempts made by malicious nodes to perceive the nature of activities and to obtain information transacted in the network without disrupting the operation. Active attacks disrupt the operation of the network. Those active attacks that are executed by nodes outside the network are called external attacks, and those that are performed by nodes belonging to the same network are called internal attacks. Nodes that perform internal attacks are compromised nodes. The major security threats that exist in ad hoc wireless networks are as follows:
• Denial of service: The attack effected by making the network resource unavailable for service to other nodes, either by consuming the bandwidth or by overloading the system, is known as denial of service (DoS). A simple scenario in which a DoS attack interrupts the operation of ad hoc wireless networks is by keeping a target node busy by making it process unnecessary packets.
• Resource consumption: The scarce availability of resources in ad hoc wireless network makes it an easy target for internal attacks, particularly aiming at consuming resources available in the network. The major types of resource-consumption attacks are the following:
– Energy depletion: Since the nodes in ad hoc wireless networks are highly constrained by the energy source, this type of attack is basically aimed at depleting the battery power of critical nodes by directing unnecessary traffic through them.
– Buffer overflow: The buffer overflow attack is carried out either by filling the routing table with unwanted routing entries or by consuming the data packet buffer space with unwanted data. Such attacks can lead to a large number of data packets being dropped, leading to the loss of critical information. Routing table attacks can lead to many problems, such as preventing a node from updating route information for important destinations and filling the routing table with routes for nonexistent destinations.
• Host impersonation: A compromised internal node can act as another node and respond with appropriate control packets to create wrong route entries, and can terminate the traffic meant for the intended destination node.
• Information disclosure: A compromised node can act as an informer by deliberate disclosure of confidential information to unauthorized nodes. Information such as the amount and the periodicity of traffic between a selected pair of nodes and pattern of traffic changes can be very valuable for military applications. The use of filler traffic (traffic generated for the sole purpose of changing the traffic pattern) may not be suitable in resource-constrained ad hoc wireless networks.
• Interference: A common attack in defense applications is to jam the wireless communication by creating a wide-spectrum noise. This can be done by using a single wide-band jammer, sweeping across the spectrum. The MAC and the physical layer technologies should be able to handle such external threats.
5.2.9 Addressing and Service Discovery
Addressing and service discovery assume significance in ad hoc wireless networks due to the absence of any centralized coordinator. An address that is globally unique in the connected part of the ad hoc wireless network is required for a node in order to participate in communication. Auto-configuration of addresses is required to allocate non-duplicate addresses to the nodes. In networks where the topology is highly dynamic, frequent partitioning and merging of network components require duplicate address-detection mechanisms in order to maintain unique addressing throughout the connected parts of the network. Nodes in the network should be able to locate services that other nodes provide. Hence efficient service advertisement mechanisms are necessary. Topological changes force a change in the location of the service provider as well, hence fixed positioning of a server providing a particular service is ruled out. Rather, identifying the current location of the service provider gathers importance. The integration of service discovery with the route-acquisition mechanism, though it violates the traditional design objectives of the routing protocol, is a viable alternative. However, provisioning of certain kinds of services demands authentication, billing, and privacy that in turn require the service discovery protocols to be separated from the network layer protocols.
5.2.10 Energy Management
Energy management is defined as the process of managing the sources and consumers of energy in a node or in the network as a whole for enhancing the lifetime of the network. Shaping the energy discharge pattern of a node's battery to enhance the battery life; finding routes that result in minimum total energy consumption in the network; using distributed scheduling schemes to improve battery life; and handling the processor and interface devices to minimize power consumption are some of the functions of energy management. Energy management can be classified into the following categories:
• Transmission power management: The power consumed by the radio frequency (RF) module of a mobile node is determined by several factors such as the state of operation, the transmission power, and the technology used for the RF circuitry. The state of operation refers to the transmit, receive, and sleep modes of the operation. The transmission power is determined by the reachability requirement of the network, the routing protocol, and the MAC protocol employed.
The RF hardware design should ensure minimum power consumption in all the three states of operation. Going to the sleep mode when not transmitting or receiving can be done by additional hardware that can wake up on reception of a control signal. Power conservation responsibility lies across the data link, network, transport, and application layers. By designing a data link layer protocol that reduces unnecessary retransmissions, by preventing collisions, by switching to standby mode or sleep mode whenever possible, and by reducing the transmit/receive switching, power management can be performed at the data link layer.
The use of a variable power MAC protocol can lead to several advantages that include energy-saving at the nodes, increase in bandwidth reuse, and reduction in interference. Also, MAC protocols for directional antennas are at their infancy. The network layer routing protocols can consider battery life and relaying load of the intermediate nodes while selecting a path so that the load can be balanced across the network, in addition to optimizing and reducing the size and frequency of control packets. At the transport layer, reducing the number of retransmissions, and recognizing and handling the reason behind the packet losses locally, can be incorporated into the protocols. At the application layer, the power consumption varies with applications. In a mobile computer, the image/video processing/playback software and 3D gaming software consume higher power than other applications. Hence application software developed for mobile computers should take into account the aspect of power consumption as well.
• Battery energy management: The battery management is aimed at extending the battery life of a node by taking advantage of its chemical properties, discharge patterns, and by the selection of a battery from a set of batteries that is available for redundancy. Recent studies showed that pulsed discharge of a battery gives longer life than continuous discharge. Controlling the charging rate and discharging rate of the battery is important in avoiding early charging to the maximum charge or full discharge below the minimum threshold. This can be achieved by means of embedded charge controllers in the battery pack. Also, the protocols at the data link layer and network layer can be designed to make use of the discharge models. Monitoring of the battery for voltage levels, remaining capacity, and temperature so that proactive actions (such as incremental powering off of certain devices, or shutting down of the mobile node when the voltage crosses a threshold) can be taken is required.
• Processor power management: The clock speed and the number of instructions executed per unit time are some of the processor parameters that affect power consumption. The CPU can be put into different power saving modes during low processing load conditions. The CPU power can be completely turned off if the machine is idle for a long time. In such cases, interrupts can be used to turn on the CPU upon detection of user interaction or other events.
• Devices power management: Intelligent device management can reduce power consumption of a mobile node significantly. This can be done by the operating system (OS) by selectively powering down interface devices that are not used or by putting devices into different power-saving modes, depending on their usage. Advanced power management features built into the operating system and application softwares for managing devices effectively are required.
5.2.11 Scalability
Even though the number of nodes in an ad hoc wireless network does not grow in the same magnitude as today's Internet, the operation of a large number of nodes in the ad hoc mode is not far away. Traditional applications such as military, emergency operations, and crowd control may not lead to such a big ad hoc wireless network. Commercial deployments of ad hoc wireless networks that include wireless mesh networks show early trends for a widespread installation of ad hoc wireless networks for mainstream wireless communication.
For example, the latency of path-finding involved with an on-demand routing protocol in a large ad hoc wireless network may be unacceptably high. Similarly, the periodic routing overhead involved in a table-driven routing protocol may consume a significant amount of bandwidth in such large networks.
Also a large ad hoc wireless network cannot be expected to be formed by homogeneous nodes, raising issues such as widely varying resource capabilities across the nodes. A hierarchical topology-based system and addressing may be more suitable for large ad hoc wireless networks. Hybrid architectures that combine the multi-hop radio relaying in the presence of infrastructure may improve scalability.
5.2.12 Deployment Considerations
The deployment of ad hoc wireless networks involves actions different from those of wired networks. It requires a good amount of planning and estimation of future traffic growth over any link in the network. The time-consuming planning stage is followed by the actual deployment of the network. The cost and time required for laying copper cables or fiber cables make it difficult to reconfigure any partial deployment that has already been done. The deployment of a commercial ad hoc wireless network has the following benefits when compared to wired networks:
• Low cost of deployment: The use of multi-hop wireless relaying essentially eliminates the requirement of laying cables and maintenance in a commercial deployment of communication infrastructure. Hence the cost involved is much lower than that of wired networks.
• Incremental deployment: In commercial wireless WANs based on ad hoc wireless networks, deployment can be performed incrementally over geographical regions of the city. The deployed part of the network starts functioning immediately after the minimum configuration is done. For example, during the deployment process for covering a highway, whenever each radio relaying equipment is installed on the highway side, it can be commissioned.
• Short deployment time: Compared to wired networks, the deployment time is considerably less due to the absence of any wired links. Also, wiring a dense urban region is extremely difficult and time-consuming in addition to the inconvenience caused.
• Reconfigurability: The cost involved in reconfiguring a wired network covering a metropolitan area network (MAN) is very high compared to that of an ad hoc wireless network covering the same service area. Also, the incremental deployment of ad hoc wireless networks might demand changes in the topology of the fixed part (e.g., the relaying devices fixed on lamp posts or rooftops) of the network at a later stage.
The issues and solutions for deployment of ad hoc wireless networks vary with the type of applications and the environment in which the networks are to be deployed. The following are the major issues to be considered in deploying an ad hoc wireless network:
• Scenario of deployment: The scenario of deployment assumes significance because the capability required for a mobile node varies with the environment in which it is used. The capabilities required for the mobile nodes that form an ad hoc wireless network among a fleet of ships are not the same as those required for forming an ad hoc wireless network among a set of notebook computers at a conference. The following are some of the different scenarios in which the deployment issues vary widely.
– Military deployment: The military deployment of an ad hoc wireless network may be data-centric (e.g., a wireless sensor network) or user-centric (e.g., soldiers or armored vehicles carrying soldiers equipped with wireless communication devices). The data-centric networks handle a different pattern of data traffic and can be partially comprised of static nodes, whereas the user-centric network consists of highly mobile nodes with or without any support from any infrastructure (e.g., military satellite constellations). The vehicle-mounted nodes have at their disposal better power sources and computational resources, whereas the hand-held devices are constrained by energy and computational resources. Thus the resource availability demands appropriate changes in the protocols employed. Also, the military environment requires secure communication. Routing should involve as few nodes as possible to avoid possible leakage of information. Flat addressing schemes are preferred to hierarchical addressing since the latter addressing requires paths to be set up through the hierarchy, and hence the chances of unreliable nodes forwarding the packets are high.
– Emergency operations deployment: This kind of application scenario demands a quick deployment of rescue personnel equipped with hand-held communication equipment. Essentially, the network should provide support for time-sensitive traffic such as voice and video. Short data messaging can be used in case the resource constraints do not permit voice communication. Also in this scenario, a flat fixed addressing scheme with a static configuration is preferred. Typically, the size of the network for such applications is not more than 100 nodes. The nodes are fully mobile without expecting support from any fixed infrastructure.
– Commercial wide-area deployment: One example of this deployment scenario is the wireless mesh networks. The aim of the deployment is to provide an alternate communication infrastructure for wireless communication in urban areas and areas where a traditional cellular base station cannot handle the traffic volume. This scenario assumes significance as it provides very low cost per bit transferred compared to the wide-area cellular network infrastructure. Another major advantage of this application is the resilience to failure of a certain number of nodes. Addressing, configuration, positioning of relaying nodes, redundancy of nodes, and power sources are the major issues in deployment. Billing, provisioning of QoS, security, and handling mobility are major issues that the service providers need to address.
– Home network deployment: The deployment of a home area network needs to consider the limited range of the devices that are to be connected by the network. Given the short transmission ranges of a few meters, it is essential to avoid network partitions. Positioning of relay nodes at certain key locations of a home area network can solve this. Also, network topology should be decided so that every node is connected through multiple neighbors for availability.
• Required longevity of network: The deployment of ad hoc wireless networks should also consider the required longevity of the network. If the network is required for a short while (e.g., the connectivity among a set of researchers at a conference and the connectivity required for coordination of a crowd control team), battery-powered mobile nodes can be used. When the connectivity is required for a longer duration of time, fixed radio relaying equipment with regenerative power sources can be deployed. A wireless mesh network with roof-top antennas deployed at a residential zone requires weather-proof packages so that the internal circuitry remains unaffected by the environmental conditions. In such an environment, the mesh connectivity is planned in such a way that the harsh atmospheric factors do not create network partitions.
• Area of coverage: In most cases, the area of coverage of an ad hoc wireless network is determined by the nature of application for which the network is set up. For example, the home area network is limited to the surroundings of a home. The mobile nodes' capabilities such as the transmission range and associated hardware, software, and power source should match the area of coverage required. In some cases where some nodes can be fixed and the network topology is partially or fully fixed, the coverage can be enhanced by means of directional antennas.
• Service availability: The availability of network service is defined as the ability of an ad hoc wireless network to provide service even with the failure of certain nodes. Availability assumes significance both in a fully mobile ad hoc wireless network used for tactical communication and in partially fixed ad hoc wireless networks used in commercial communication infrastructure such as wireless mesh networks. In the case of wireless mesh networks, the fixed nodes need to be placed in such a way that the failure of multiple nodes does not lead to lack of service in that area. In such cases, redundant inactive radio relaying devices can be placed in such a way that on the event of failure of an active relaying node, the redundant relaying device can take over its responsibilities.
• Operational integration with other infrastructure: Operational integration of ad hoc wireless networks with other infrastructures can be considered for improving the performance or gathering additional information, or for providing better QoS. In the military environment, integration of ad hoc wireless networks with satellite networks or unmanned aerial vehicles (UAVs) improves the capability of the ad hoc wireless networks. Several routing protocols assume the availability of the global positioning system (GPS), which is a satellite-based infrastructure by which the geographical location information can be obtained as a resource for network synchronization and geographical positioning. In the commercial world, the wireless mesh networks that service a given urban region can interoperate with the wide-area cellular infrastructure in order to provide better QoS and smooth handoffs across the networks. Handoff to a different network can be done in order to avoid call drops when a mobile node with an active call moves into a region where service is not provided by the current network.
• Choice of protocols: The choice of protocols at different layers of the protocol stack is to be done taking into consideration the deployment scenario. A TDMA-based and insecure MAC protocol may not be the best suited compared to a CDMA-based MAC protocol for a military application. The MAC protocol should ensure provisioning of security at the link level. At the network layer, the routing protocol has to be selected with care. A routing protocol that uses geographical information (GPS information) may not work well in situations where such information is not available. For example, the search-and-rescue operation teams that work in extreme terrains or underground or inside a building may not be able to use such a routing protocol. An ad hoc wireless network with nodes that cannot have their power sources replenished should use a routing protocol that does not employ periodic bea-cons for routing. The periodic beacons, or routing updates, drain the battery with time. In situations of high mobility, for example, an ad hoc wireless network formed by devices connected to military vehicles, the power consumption may not be very important and hence one can employ beacon-based routing protocols for them. The updated information about connectivity leads to improved performance. In the case of deployment of wireless mesh networks, the protocols should make use of the fixed nodes to avoid unstable paths due to the mobility of the relaying nodes.
At the transport layer, the connection-oriented or connectionless protocol should be adapted to work in the environment in which the ad hoc wireless network is deployed. In a high-mobility environment, path breaks, network partitions, and remerging of partitions are to be considered, and appropriate actions should be taken at the higher layers. This can be extended to connectionless transport protocols to avoid congestion. Also, packet loss arising out of congestion, channel error, link break, and network partition is to be handled differently in different applications. The timer values at different layers of the protocol stack should be adapted to the deployment scenarios.
5.3 AD HOC WIRELESS INTERNET
Similar to the wireless Internet discussed in Chapter 4, the ad hoc wireless Internet extends the services of the Internet to the end users over an ad hoc wireless network. Some of the applications of the ad hoc wireless Internet are wireless mesh networks, provisioning of temporary Internet services to major conference venues, sports venues, temporary military settlements, battlefields, and broadband Internet services in rural regions. A schematic diagram of the ad hoc wireless Internet is shown in Figure 5.7.
Figure 5.7. A schematic diagram of the ad hoc wireless Internet.
The major issues to be considered for a successful ad hoc wireless Internet are the following:
• Gateways: Gateway nodes in the ad hoc wireless Internet are the entry points to the wired Internet. The major part of the service provisioning lies with the gateway nodes. Generally owned and operated by a service provider, gateways perform the following tasks: keeping track of the end users, bandwidth management, load balancing, traffic shaping, packet filtering, bandwidth fairness, and address, service, and location discovery.
• Address mobility: Similar to the Mobile IP discussed in Chapter 4, the ad hoc wireless Internet also faces the challenge of address mobility. This problem is worse here as the nodes operate over multiple wireless hops. Solutions such as Mobile IP can provide temporary alternatives for this.
• Routing: Routing is a major problem in the ad hoc wireless Internet, due to the dynamic topological changes, the presence of gateways, multi-hop relaying, and the hybrid character of the network. The possible solution for this is the use of a separate routing protocol, which is discussed in Chapter 7, for the wireless part of the ad hoc wireless Internet. Routing protocols discussed in Chapter 13 are more suitable as they exploit the presence of gateway nodes.
• Transport layer protocol: Even though several solutions for transport layer protocols exist for ad hoc wireless networks, unlike other layers, the choice lies in favor of TCP's extensions proposed for ad hoc wireless networks. Split approaches that use traditional wired TCP for the wired part and a specialized transport layer protocol for the ad hoc wireless network part can also be considered where the gateways act as the intermediate nodes at which the connections are split. Several factors are to be considered here, the major one being the state maintenance overhead at the gateway nodes.
• Load balancing: It is likely that the ad hoc wireless Internet gateways experience heavy traffic. Hence the gateways can be saturated much earlier than other nodes in the network. Load balancing techniques are essential to distribute the load so as to avoid the situation where the gateway nodes become bottleneck nodes. Gateway selection strategies and load balancing schemes discussed in Chapter 13 can be used for this purpose.
• Pricing/billing: Since Internet bandwidth is expensive, it becomes very important to introduce pricing/billing strategies for the ad hoc wireless Internet. Gateway is the preferred choice for charging the traffic to and from the Internet. Pricing schemes discussed in Chapter 13 can be used for this purpose. A much more complex case is pricing the local traffic (traffic within the wireless part, that is, it originated and terminated within the wireless part without passing through the gateway nodes), where it becomes necessary to have a dedicated, secure, and lightweight pricing/billing infrastructure installed at every node.
• Provisioning of security: The inherent broadcast nature of the wireless medium attracts not just the mobility seekers but also potential hackers who love to snoop on important information sent unprotected over the air. Hence security is a prime concern in the ad hoc wireless Internet. Since the end users can utilize the ad hoc wireless Internet infrastructure to make e-commerce transactions, it is important to include security mechanisms in the ad hoc wireless Internet.
• QoS support: With the widespread use of voice over IP (VoIP) and growing multimedia applications over the Internet, provisioning of QoS support in the ad hoc wireless Internet becomes a very important issue. As discussed in Chapter 10, this is a challenging problem in the wired part as well as in the wireless part.
• Service, address, and location discovery: Service discovery in any network refers to the activity of discovering or identifying the party which provides a particular service or resource. In wired networks, service location protocols exist to do the same, and similar systems need to be extended to operate in the ad hoc wireless Internet as well. Address discovery refers to the services such as those provided by address resolution protocol (ARP) or domain name service (DNS) operating within the wireless domain. Location discovery refers to different activities such as detecting the location of a particular mobile node in the network or detecting the geographical location of nodes. Location discovery services can provide enhanced services such as routing of packets, location-based services, and selective region-wide broadcasts.
Figure 5.8 shows a wireless mesh network that connects several houses to the Internet through a gateway node. Such networks can provide highly reliable broadband wireless networks for the urban as well as the rural population in a cost-effective manner with fast deployment and reconfiguration. This wireless mesh network is a special case of the ad hoc wireless Internet where mobility of nodes is not a major concern as most relay stations and end users use fixed transceivers. Figure 5.8 shows that house A is connected to the Internet over multiple paths (path 1 and path 2).
Figure 5.8. An illustration of the ad hoc wireless Internet implemented by a wireless mesh network.
5.4 SUMMARY
In this chapter, the major issues and applications of ad hoc wireless networks were described. The design issues for every layer of protocol stack and deployment scenarios were discussed. The applications of ad hoc wireless networks include military applications, collaborative and distributed computing, emergency operations, and hybrid wireless architectures. The important deployment issues for ad hoc wireless networks are scenario of deployment, required longevity of network, area of coverage, service availability, operational integration with other infrastructure, and choice of protocols. Each of the deployment issues has been discussed in detail. The ad hoc wireless Internet was another important topic discussed in this chapter, in which the objectives, challenges, and the application scenarios were discussed.
5.5 PROBLEMS
- Describe a common method used in alleviating the hidden terminal problem at the MAC layer.
- How is the loop-free property ensured in on-demand routing protocols? (Hint: On-demand routing protocols do not maintain network topology information and obtain the necessary path as and when required by using a connection-establishment process.)
- How is the loop-free property ensured in table-driven routing protocols? (Hint: In table-driven routing protocols, nodes maintain the network topology information in the form of routing tables by periodically exchanging routing information.)
- Identify and elaborate some of the important issues in pricing for multi-hop wireless communication.
- What is replay attack? How can it be prevented?
- Why is power management important for ad hoc wireless networks?
- What are the trade-offs to be considered in the design of power management schemes?
- What role does the routing protocol play in the provisioning of QoS guarantees for ad hoc wireless networks?
- What kind of multiple access technology is suitable in (a) a military ad hoc network environment, and (b) a home ad hoc network environment?
- Referring to Figure 5.5, a mobile station (node) moving at high speeds would experience frequent handoffs. Discuss the pros and cons of using MobileIP for address and service discovery.
- Referring to Figure 5.5, what is the rate of handoffs that a mobile node experiences when it moves at a speed of 144 Km/hour, assuming that the relay node's transmission range (radius) is 500 m.
- Discuss the pros and cons of a routing protocol that uses GPS information for an ad hoc wireless network for search-and-rescue operations.
- List the major advantages of the ad hoc wireless Internet.
BIBLIOGRAPHY
[1] "DARPA Home Page," http://www.darpa.mil
[2] "IETF MANET Working Group Information," http://www.ietf.org/html.charters/manet-charter.html
[3] Y. D. Lin and Y. C. Hsu, "Multi-Hop Cellular: A New Architecture for Wireless Communications," Proceedings of IEEE INFOCOM 2000, pp. 1273-1282, March 2000.
[4] A. N. Zadeh, B. Jabbari, R. Pickholtz, and B. Vojcic, "Self-Organizing Packet Radio Ad Hoc Networks with Overlay," IEEE Communications Magazine, vol. 40, no. 6, pp. 140-157, June 2002.
[5] R. Ananthapadmanabha, B. S. Manoj, and C. Siva Ram Murthy, "Multi-Hop Cellular Networks: The Architecture and Routing Protocol," Proceedings of IEEE PIMRC 2001, vol. 2, pp. 78-82, October 2001.
[6] H. Wu, C. Qiao, S. De, and O. Tonguz, "Integrated Cellular and Ad Hoc Relaying Systems: iCAR," IEEE Journal on Selected Areas in Communications, vol. 19, no. 10, pp. 2105-2115, October 2001.
[7] V. D. Park and M. S. Corson, "A Highly Adaptive Distributed Routing Algorithm for Mobile Wireless Networks," Proceedings of IEEE INFOCOM 1997, pp. 1405-1413, April 1997.
[8] S. J. Lee and M. Gerla, "SMR: Split Multipath Routing with Maximally Disjoint Paths in Ad Hoc Networks," Proceedings of IEEE ICC 2001, vol. 10, pp. 3201-3205, June 2001.
[9] J. Raju and J. J. Garcia-Luna-Aceves, "A New Approach to On-Demand Loop-Free Multipath Routing," Proceedings of IEEE ICCCN 1999, pp. 522-527, October 1999.
[10] I. F. Akyildiz, W. Su, Y. Sankarasubramaniam, and E. Cayirci, "A Survey on Sensor Networks," IEEE Communications Magazine, vol. 40, no. 8, pp 102-114, August 2002.
[11] D. Christo Frank, B. S. Manoj, and C. Siva Ram Murthy, "Throughput Enhanced Wireless in Local Loop (TWiLL) — The Architecture, Protocols, and Pricing Schemes," Proceedings of IEEE LCN 2002, pp. 177-186, November 2002.
[12] H. Deng, W. Li, and D. P. Agrawal, "Routing Security in Wireless Ad Hoc Networks," IEEE Communications Magazine, vol. 40, no. 10, pp. 70-75, October 2002.
[13] C. L. Fullmer and J. J. Garcia-Luna-Aceves, "Solutions to Hidden Terminal Problems in Wireless Networks," Proceedings of ACM SIGCOMM 1997, pp. 39-49, September 1997.
[14] Z. J. Haas, "The Routing Algorithm for the Reconfigurable Wireless Networks," Proceedings of IEEE ICUPC 1997, vol. 2, pp. 562-566, October 1997.
[15] S. B. Lee, A. Gahng-seop, X. Zhang, and A. T. Campbell, "INSIGNIA: An IP-Based Quality of Service Framework for Mobile Ad Hoc Networks," Journal of Parallel and Distributed Computing, vol. 60, no. 4, pp. 374-406, April 2000.
[16] C. R. Lin and M. Gerla, "Asynchronous Multimedia Multi-Hop Wireless Networks," Proceedings of IEEE INFOCOM 1997, vol. 1, pp. 118-125, April 1997.
[17] J. Liu and S. Singh, "ATCP: TCP for Mobile Ad Hoc Networks," IEEE Journal on Selected Areas in Communications, vol. 19, no. 7, pp. 1300-1315, July 2001.
[18] R. Sivakumar, P. Sinha, and V. Bharghavan, "CEDAR: A Core-Extraction Distributed Ad Hoc Routing Algorithm," IEEE Journal on Selected Areas in Communications, vol. 17, no. 8, pp. 1454-1465, August 1999.
[19] L. Subramanian and R. H. Katz, "An Architecture for Building Self-Configurable Systems," Proceedings of ACM MOBIHOC 2000, pp. 63-73, August 2000.
[20] N. H. Vaidya, "Weak Duplicate Address Detection in Mobile Ad Hoc Networks," Proceedings of ACM MOBIHOC 2002, pp. 206-216, June 2002.
[21] B. S. Manoj and C. Siva Ram Murthy, "Ad Hoc Wireless Networks: Issues and Challenges," Technical Report, Department of Computer Science and Engineering, Indian Institute of Technology, Madras, India, November 2003.