Misbehaviour Detection for Wireless Sensor Networks - Necessary or Not?

As radio communication modules combined with adequate hardware based cryptography are becoming available at low cost, sensor network security has reached a new level. However it is still possible to gain access to networks, even if they require authorisation and authentication, as single nodes cannot be assumed to be tamper proof. Therefore the question remains what kind of misbehaviour is possible within a secured network and how it can be detected. A possible solution to detect misbehaviour is the usage of an artificial immune system (AIS). The advantage of an AIS in contrast to protocol based security improvements is the generality of the approach, as it tries to identify anomalies which are not known to exist within the regular range of network behaviour. In this paper we present some results obtained from our simulation experiments looking at the possibilities of artificial immune system based misbehaviour detection.