Inspiration from genetics to promote recognition and protection within ad hoc sensor networks

Abstract: This work illustrates potentials for recognition within {\em ad hoc} sensor networks if their nodes possess individual inter-related biologically inspired genetic codes. The work takes ideas from natural immune systems protecting organisms from infection. Nodes in the present proposal have individual gene sets fitting into a self organised phylogenetic tree. Members of this population are genetically ''relatives''. Outsiders cannot easily copy or introduce a new node in the network without going through a process of conception between two nodes in the population. Related nodes can locally decide to check each other for their genetic relation without directly revealing their gene sets. A copy/clone of a gene sequence or a random gene set will appear as alien. Nodes go through a cycle of introduction (conception or ''birth'') with parents in the network and later exit from it (''death''). Hence the phylogenetic tree is dynamic or possesses a genetic drift. Typical lifetimes of gene sets and number of offspring from different parents affect this genetic drift and the level of correlation between gene sets. The frequency of mutations similarly affects the gene pool. Correlation between genes of the nodes implies a common secret for cryptographic material for communication and consistency check facilitating intrusion detection and tracing of events. A node can by itself (non-specifically) recognise an adversary if it does not respond properly according to its genes. Nodes can also collaborate to recognise adversaries by communicating response from intruders to check for consistency with the whole gene pool (phylogenetic tree).