Introduction

Ensuring the security of both the collected data and the process of data collection is vital for the success of WSNs. Because of the constraints of the particular applications and the resource limitations, the security of WSNs is vastly different from that of conventional wired networks. For the example of military applications, wireless sensor nodes usually are sent to an unattended environment (e.g., the battlefield). In these scenarios, wireless sensor nodes are easier to capture or destroy. Thus, the foremost important thing for WSNs is that they can tolerate the dysfunction of a certain number of nodes; e.g., the network formed by nondestroyed nodes should be always connected. The second possible attack for WSNs is that the enemy could distribute a certain number of faked sensor nodes to disturb or even disrupt the communications of legitimate sensors. It is important for a sensor network to design a security mechanism to protect the sensor nodes from malicious attack or to ensure that the sensor network can “tolerate” the malicious attack to some extent. The second challenge for the design of security mechanisms for WSNs is that sensor nodes are always equipped with limited battery and memory. Thus, the traditional public-key–based schemes, such as the Rivest–Shamir–Adleman (RSA) and Diffie–Hellman (D-H) protocols, are not suitable for WSNs. For example, Mica Mote, produced by UC Berkeley, has 128 kb Flash and 4 kb RAM.