Effectiveness of DCFM on different mobility models

Abstract

Simulation based network protocol testing requires that one model reality correctly. Mis-modelling can render results invalid. Hence, one must determine the context in which a solution is to be implemented for the results to be meaningful. In this work, we investigate whether the security augmentation technique DCFM (denial contradictions with fictitious nodes) for OLSR based mobile ad-hoc networks (MANET) is valid for multiple contexts. We accomplish this by running the node isolation attack and DCFM defense on a variety of mobility models, comparing the results. We show that for most models the technique performs well, while for a specific scenario (e.g. street based) it fails. Additionally we describe the infrastructure that was built to support the street based simulations.