Reference Architecture of an Autonomous Agent for Cyber Defense of Complex Military Systems
Résumé
Military strategies will shortly make intensive use of autonomous systems while the Internet of Battle Things (IoBT) will grow military systems’ complexity to new heights. The cyber defense of the battlespace will then become arduous for humans, if not impossible, due to disconnections, the difficulty of supervising masses of interconnected devices, and the scarcity of cyber defense competences on the battleground. An autonomous intelligent cyber defense of the battlefield becomes necessary in such a context. In response to such needs, this chapter presents and illustrates the rationale, concept and future research directions of (Multiple) Autonomous Intelligent Cyber defense Agents, (M)AICA, and NATO’s initial AICA Reference Architecture, AICARA.
Mots clés
Cyber security
adaptive cyber defense
cyber attacks
cyber defense
network security
autonomous systems
science of security
game theory
Autonomic computing
Cyber deceptionmachine learning
programming languages
graph theory
Adversarial artificial intelligence
Generative adversarial networks
Quantitative performance metrics
Deep learning