Performance evaluation of Wireless Sensor Networks (WSNs), like any communications network system, can be conducted by using simulation analysis, analytic modeling, and measurement/testing techniques. Evaluation of WSN systems is needed at every stage in their life. There is no point in designing and implementing a new system that does not have competitive performance, and does not meet the objectives and performance evaluation and quality of service requirements. Performance evaluation of an existing system is also important since it helps to determine how well it is performing and whether any improvements are needed in order to enhance the performance [1].

After a system has been built and is running, its performance can be assessed by using the measurement technique. In order to evaluate the performance of a component or subsystem that cannot be measured, for example, during the design and development phases, it is necessary to use analytic and/or simulation modeling so as to predict the performance [1–46].

The objective of this chapter is to provide an up-to-date treatment of the techniques that can be used to evaluate the performance of WSN systems.

Background information

Wireless sensor networks (WSNs) are unique in certain aspects that make them different from other wireless networks. These aspects include:

Vehicular networking, the exchange of information in the car and between cars, has been on the mind of researchers since at least the often-cited 1939 New York World's Fair. Here, in its Futurama exhibit, General Motors revealed utopian visions of what highways and cities might look like twenty years later. In fact, many of the visions of intelligent transportation systems (ITSs) showcased there, as well as in the exhibit designer's 1940 book Magic Motorways (Bel Geddes 1940), such as that “car-to-car radio hook-up might be used to advise a driver nearing an intersection of the approach of another car or even to maintain control of speed and spacing of cars in the same traffic lane”, are still being pursued today. Modern vehicles collect huge amounts of information from on-board sensors, and this information is made available to the in-car network and ready for sharing with other cars – not just for the described visions of intersection assistance systems and platooning, i.e., road-train applications, but also for a whole wealth of new applications. Today, with in-car networks merging into networks of cars, these early visions seem closer to reality than ever.

But why did we have to wait this long?

Hugely many research projects have been undertaken since Magic Motorways was written, all of which tried to make visions of ITS a reality (Jurgen 1991). Among the most notable of research initiatives were the Japanese CACS, US ERGS, and European ALI projects for urban route guidance in the late 1960s to late 1970s, the European Prometheus project for autonomous driving (1986–1995), and the US PATH project for cooperative driving (1986–1992). Evidently, the majority of these initiatives led to working prototypes and successful field operational tests; yet, commercial success failed to match the projects' promises.

A possible explanation is given by Chen & Ervin (1990): early approaches were simply too visionary for their time, commonly focusing on infrastructure-less solutions, which could not be supported by current technology. The 1980s then saw a shift of attention from the more long-term goals of complete highway automation to nearer-term goals such as driver-advisory functions.

Providing security to wireless sensor networks is very challenging, as they include protection against damages, losses, attacks, and dangers. Moreover, a wireless sensor node has limited computation power, limited memory, and limited I/O resources. The classic security issues that are usually considered in wireless sensor networks are upholding the secrecy and avoiding intrusion. Securing access to wireless networks in general is a difficult task when compared to fixed/wired networks because wireless networks use wireless transmission medium. Securing access to WSNs is more challenging than for traditional wireless networks. This is mainly due to the limited resources of WSNs and to the harsh working environments of these systems in most cases.

In this chapter, we present key issues, challenges, vulnerabilities, attacks, existing solutions, and comparison of major security techniques related to WSNs.

Background

In general, WSNs are heterogeneous systems. They contain general-purpose computing elements with actuators and tiny sensors. Moreover, these computing elements have limited computational power, limited power, limited bandwidth, and limited peripherals. These aspects of WSNs make it difficult and challenging to design a secure WSN system [1–71], as secured schemes require computational power, large memory, and more power consumption, among other resources.

Moreover, providing security in WSNs is not an easy task because of the resource limitation on sensor nodes, high risk of physical attacks, density and size of networks, unknown topology prior to deployment, and also due to the nature and characteristics of wireless communication channels.

Electronics are playing an ever-increasing role in today's vehicles. They have gone from humble beginnings in the 1970s that saw features like electronic fuel injection and power door locks become commonplace or the mass-market introduction of anti-lock braking systems in the 1980s to today's 3 km of wiring and 50 kg of electrical systems operating in a typical modern car. With electrical systems contributing to at least one third of today's breakdowns, it is clear that modern cars just could not be driven without electronics.

The use cases that most readily come to mind might be engine control, or more recent advances in chassis electrification such as an anti-lock braking system (ABS) or an electronic stability program (ESP), maybe also modern infotainment systems like DVD players streaming video to the rear seats, navigation systems, or video cameras. Yet, modern cars contain a vast number of small electronic subsystems – from windshield wipers, door locks, power windows, and adaptive lights to seat and mirror adjustment, climate control, and dashboard displays. All these devices are controlled by electronic control units (ECUs).

Early on, this motivated the introduction of a way to diagnose faults in the electronics system, requiring digital data communication between embedded systems and diagnostic tools. Such systems were first designed for use exclusively on the vehicle assembly line and were very specific to not just the manufacturer but also the component being diagnosed. One of the earliest examples is the assembly line diagnostic link (ALDL) system used in the early 1980s for engine control module diagnosis.

The need for a digital link between ECUs and other components was further driven by the increasing complexity of connecting individual electric systems. In the early 1980s this was still easily possible using dedicated wires for each interconnection, or even for each task, but it became very clear that this would not much longer be the case; and rightly so: Today, switching on the left-turn signal involves no fewer than eight individual embedded systems.

With the increasing engagement of the major industry players to bring inter-vehicle communication (IVC) onto the market, the need to secure the communication between vehicles and the infrastructure became an important issue. Apart from satisfying the demand for deploying closed-market systems that offer services only to paying customers, security is also necessary in order to prevent fraud and malicious attacks. Just recently, it has been demonstrated that IVC-related systems are not as secure as necessary: Attackers successfully took over control of car electronics via tire-pressure-measurement-systems or they attacked electronic message boards along a highway. Even spoofed traffic-information transmissions via TMC have been demonstrated. In this chapter, we study possible security solutions for vehicular networks, focusing on their practical relevance. We review not only the generic security primitives but also their applicability and limits. Furthermore, we look into the very critical balance between security and privacy: The more secure a system is made, the more severely the driver's privacy is impacted. Therefore, we also investigate location privacy and outline how the driver's privacy can be increased. In particular, we investigate the use of pseudonyms, time-varying pseudonym pools, and the exchange of pseudonyms.

This chapter is organized as follows.

1. • Security primitives (Section 7.1) – In this section, we briefly review the general security objectives before investigating the specific security relationships relevant for vehicular networks. The main focus is on introducing the concept of certificates and their use for digitally signing messages. We also investigate the fundamental relationship between security and privacy.

2. • Securing vehicular networks (Section 7.2) – This is the key section on enabling security in vehicular networks. The key technology proposed is the use of certificates for digitally signing messages such as periodic CAMs. We further investigate how the resulting performance issues can be solved as well as how certificates can be revoked if keys have been compromised. We conclude this section with a brief overview on using context information such as geographic position to increase security.

3. […]

Underwater sensor networks (UWSNs) are wireless networks of autonomous sensor-aided devices, called motes or sensor nodes, deployed over a region of water for the collaborative execution of a given task. Nearly 70% of the earth’s surface is covered by water, mainly oceans. The vast majority of this area remains unexplored. The advent of UWSNs provides a new direction in the field of oceanic exploration and information collection. Major applications of UWSNs exist in both the military and civilian fields. Oceanographic data collection, environmental monitoring, pollution monitoring and control, intrusion detection, mapping of underwater area, detection of explosives, mines, oil and minerals, and guided navigation of rescue teams by collaboration with autonomous underwater vehicles (AUVs) and remotely operated vehicles (ROVs) are a few such potential applications [1].

Recently, a lot of real-world short-term deployments have been performed and long-term projects have been undertaken using UWSNs for various applications [2]. One such initial experiment was done in Seaweb [3]. Seaweb was targeted for military applications with a goal of designing specific protocols for detection of submarines and communication between them. In this case, UWSN was deployed in a coastal area, and experiments were carried out for several days. Various institutes have taken such initiatives for designing autonomous and robotic vehicles to be used in underwater exploration. In an underwater data-collection experiment undertaken by Massachusetts Institute of Technology (MIT) and Australia’s Commonwealth Scientific and Industrial Research Organisation, both fixed nodes and autonomous vehicles were used [4]. Another recent initiative was undertaken by IBM and Beacon Institute jointly [5]. The project concerns on environmental monitoring application to study and collect the biological, chemical, and physical information of the Hudson River in New York.

The intensive use of networked embedded systems is one of the key success factors in the automotive industry, also triggering a massive shortening of innovation cycles. Hundreds of so-called electronic control units (ECUs), connected by kilometers of electrical wiring, operate in today's modern car, enabling a huge variety of new functionalities ranging from safety to comfort applications. All this functionality can be realized only if the ECUs are able to communicate and to cooperate using a real-time enabled communication network in the car.

Today we are at the verge of another leap forward: This in-car network is being extended to not only connect local ECUs but also to connect the whole car to other cars and its environment using inter-vehicle communication (IVC). Relying on existing wireless Internet access using cellular networks of the third (3G) or fourth generation (4G), or novel networking technologies that are being designed specifically for use in the vehicular context such as IEEE WAVE, ETSI ITS-G5, and the IEEE 802.11p protocol, it becomes possible to use spontaneous connections between vehicles to exchange information, promising novel and sometimes futuristic applications.

Using such IVC, safety-relevant information can be exchanged that could not have been obtained using local sensors, enabling a driver to virtually see traffic through large trucks or buildings. This new idea of networked vehicles creates opportunities to not only increase road traffic safety but also improve our driving experience. Traffic jams can be prevented altogether (or at least we would be informed of jams well in advance) – and we might even be able to enable the driver to enjoying fully automated rides in a train-like convoy of cooperating vehicles on the road.

Vehicular networking, the fusion of vehicles' networks to exchange information, is the common basis on which all of these visions build.

Being fascinated with all the opportunities and challenges related to vehicular networking, we have been a part of this research community for close to ten years.

Overview and goals

Small low-cost devices powered with wireless communication technologies along with the sensing capabilities are instrumental in the inception of wireless sensor networks (WSNs). Recent years have witnessed a sharp growth in research in the area of WSNs. The characteristics of such distributed networks of sensors are that they have the potential for use in various applications in both the civilian and military fields. Enemy intrusion detection in the battlefield, object tracking, habitat monitoring, patient monitoring, and fire detection are some of the numerous potential applications of sensor networks. The ability of an infrastructure-less network setup with minimal reliance on network planning, and the ability of the deployed nodes to self-organize and self-configure without the association of any centralized control are the smart features of these networks. Leveraging the advantages of these features, the network setup is swift in challenging scenarios such as emergency, rescue, or relief operations. The smart features also enable continuous operation of the network without any intervention in case of any failure.

Along with the above-mentioned attractive features possessed by sensor networks, there are several challenges which hinder hassle-free, autonomous, and involuntary operation of these networks. Some of the challenges are attributed to issues relating to scalability, quality-of-service (QoS), energy efficiency, and security. The protocols should be light-weight enough to be suitable for these networks, which consist of small-sized sensor nodes with limited computation power. Sensor networks are often deployed in large-scale and are expected to function through years. Clearly, battery power is an issue in such cases, and can be achieved with the help of energy-efficient or energy-aware protocols. Finally, QoS is also an issue for applications which demand prompt responses.