Securing Agent-Based Security Systems

This effort advances agent reliability by addressing the problem of information security (confidentiality, integrity, and availability) in multi-agent systems. ORINCON proposes to demonstrate new technologies that are widely applicable and independent of particular agent control strategies. Our solution is to build cryptographic security shells around the agents and their communication channels, and to control access to resources where most conflicts occur. We will design an integrated environment in which security and conflict management features can be synthesized and distributed efficiently, with capability for creation of robust agent software specifically designed to protect networked agent-based systems. Protection will be achieved through an integrated approach: (1) multiple layers of security assurance, capable of evolving with time, (2) unique measures tailored to specific subversion methodologies, (3) real-time conflict management, and (4) security system self-protection. The architecture will balance additional computational burdens against current policy, scope of threat, and resource conflict potential. ORINCON brings extensive experience as a developer of agent-based systems and distributed agent security systems for Information Warfare. Initially we will design experiments to demonstrate this technology in a controlled environment. In Phase II we will implement the technology in an agent-based Information Warfare security system to clearly show its operational utility.