A Sense of Danger: Dendritic Cells Inspired Artificial Immune System (AIS) for MANET Security

AIS-based anomaly detection systems classically utilize the paradigm of self/non-self discrimination. In this approach, an algorithm learns self during a learning phase, therefore, such algorithms do not have the ability to cope with scenarios in which self is continuously changing with time. This situation is encountered once malicious nodes are to be detected in a Mobile Ad Hoc Network (MANET). Consequently, it becomes a challenge to differentiate a valid route change due to mobility from an illegal one due to tampering of routing information by malicious nodes. In this paper, we propose a dendritic cell based distributed misbehavior detection system, BeeAIS-DC, for a Bio/Nature inspired MANET routing protocol, BeeAdHoc. Our proposed system inspires from the danger theory and models the function and behavior of dendritic cells to detect the presence or absence of danger and provides a tolerogenic or immunogenic response. The proposed detection system is implemented in a well-known ns-2 simulator. Our results indicate that our detection system not only enables BeeAIS-DC to dynamically adapt its detector set to cater for a changing self due to mobility of nodes, but also is robust enough to provide significantly smaller false positives as compared to self/non-self based AIS. Moreover, the danger theory related overhead of BeeAIS-DC is minimal, and as a result, it does not degrade traditional performance metrics of BeeAdHoc. This behavior is vital for battery/bandwidth constrained mobile nodes.