Poster: Attack Survivability Prediction

Survivability analysis focuses on the ability of network entities to function during incidents such as attacks. Currently, testing survivability of mobile ad hoc networks consists of running scenarios with several configurations, often thousands, to obtain an understanding of the impacts of an attack. This process is very latent, choice of configurations are subjective or random, and results do not generalize to different scenarios. Focusing on these problems, our work-in-progress is towards a previously unexplored field of research: efficient attack survivability analysis via machine learning and an attacker-centric network representation. Using a collected dataset, we provide some evidence showing that the network representation is suitable for creating an attack survivability predictor.