A Compromise-resilient Pair-wise Rekeying Protocol in Hierarchical Wireless Sensor Networks 315 A Compromise-resilient Pair-wise Rekeying Protocol in Hierarchical Wireless Sensor Networks

Wireless sensor networks (WSNs) have been envisioned to be very useful for a broad spectrum of emerging civil and military applications (Akyildiz et al., 2002). However, sensor networks are also confronted with many security threats such as node compromise, routing disruption and false data injection, because they are normally operated in an unattended, harsh or hostile environment. Among all these threats, the WSNs are particularly vulnerable to the node compromise because sensor nodes are not tamper-proof devices. When a sensor network is deployed in unattended and hostile environments such as battlefield, the adversaries may capture and reprogram some sensor nodes, or inject some malicious sensor nodes into the network and make the network accept them as legitimate nodes. After getting control of a few nodes, the adversary can mount various attacks from inside the network (Zhang et al., 2008). Therefore, it is desirable to design key distribution protocols to support secure and robust pair-wise communication among any pair of sensors. This is a challenging task in sensor networks because they have scarce resources in energy, computation and communication. As a result, the conventional asymmetric key cryptosystem, such as RSA (Rivest et al., 1978) and Diffie-Hellman (Diffie & Hellman, 1976), can not be implemented in sensor nodes due to their very limited capacities and only lightweight energy efficient key distribution mechanisms are affordable. Furthermore, sensor nodes are low-cost and they cannot afford tamper-resistance hardware. Recent advances in physical attack show that even memory chips with built-in tamper-resistance mechanisms are subject to various memory read-out attacks. Thus, an adversary might easily capture the sensor devices to acquire their sensitive data and keys and then abuse them to further compromise the communication between other non-captured nodes. In order to conquer such node capture attack (NCA) problem, it is desirable to design protocols to support secure and robust pair-wise communication among any pair of sensors. To defend against such attack, the security mechanisms in WSNs are required. Most of existing key management schemes focus on the efficiency of bootstrapping session keys which has been intensively studied in the literature of WSNs (Cheng & Agrawal, 2005; Du et al., 2003; Eschenauer & Gligor, 2002). Traditionally, once such key system is adopted, the whole security system is established and fixed. However, when the WSN runs for a long time using a 18