Adaptive Security and Trust

In this paper we discuss ongoing work in an ambitious DARPA funded effort to develop new ways of achieving cyber security. The broad approach taken by the project is for the network to be self-aware and to self-adapt in order to dodge attacks. After all attempts to avoid attack have been applied, some attacks will still get through. In critical systems, it is not always the best or practical thing to do to shut down the network under attack. In this paper we describe how trust in cyber assets can be modeled and used to support self-adaptive techniques in continuing to operate important computational functions in a network that is actively under attack. The paper, in particular describes the trust modeling system that attempts to maintain a model of trust for networked resources using a combination of two basic ideas: Conditional trust (based on conditional preference (CP-Nets) and the principle of maximum entropy (PME).