Statistical model checking of ad hoc routing protocols in lossy grid networks

Abstract

We extend recent work by Höfner and McIver con the performances of the ad hoc routing protocols AODV and DYMO in terms of routes established. Höfner and McIver apply statistical model checking to show that on arbitrary small networks (up to 5 nodes) the most recent, and apparently more robust, DYMO protocol is less efficient than AODV. Here, we reformulate their experiments on 4x3 toroidal networks, with possibly lossy communication. As a main result we demonstrate that, in this more realistic scenario, DYMO performs significantly better than AODV.