Safari Books Online is a digital library providing on-demand subscription access to thousands of learning resources.
CHAPTER 10 Defending Against Physical Attacks in Wireless Sensor Networks accurately and completely designers consider potential threats at design time, the resources available for design, construction, testing, cost effectiveness, and attacker abilities among others. Although such approaches are obviously wel- come, there cannot be a panacea for defenses against physical attacks. We have to design effec- tive defense mechanisms against physical attacks that can be carried out by small-sized, power con- strained sensors themselves, which is the focus of this chapter. However, approaches aiming to design a degree of physically hardened sensors are important, and they can further comple- ment our orthogonal approaches presented here when defending sensor networks against physi- cal attacks. Other traditional physical defenses include camouflaging or hiding nodes and adding redundancy to the network [10]. This type of defense is particularly well suited for sensor net- works, because a constellation of many cheap 255 the desired network operations. To this end, Ref. [22] proposes a localized algorithm that can detect many kinds of misbehaving nodes, as long as the "abnormal behavior" can be mod- eled by real numbers. In Ref. [4], the authors use their lightweight authenticated broadcast proto- col to reestablish routes around failed nodes. The authors in Ref. [23] use the TESLA protocol developed [4] to allow the base station to iden- tify and isolate misbehaving nodes. Koushanafar et al. take a different approach to node failures by proposing resources, such as computing, storage, communication, sensing, and actuating, present in the network be used to back up each other [24]. For example, they propose that when the computational resources are limited, but commu- nication resources are available, nodes can send more information to reduce the need for com- putational procedures. Other approaches include keeping track of failed nodes at the base sta-