Decentralised Soft-Security in Distributed Systems
Résumé
We present an approach to intrusion detection in wireless peer-to peer networks by combining abstractions of game theory and artificial immune systems. The application to wireless networks of each of these abstractions in isolation has been well-explored, but their combined application is novel. Current detection methodology relies on a binary response, but it is hypothesised that a graded approach to response is more applicable within peer to peer networks, as nodes may be non-malicious but can exhibit varying degrees of unreliability and implied uncertainty in their operation. A rigorous simulation methodology is presented using several well researched attacks such as wormholes.