Software Tools for Information Assured Intelligent Agents (IA)^2

Software Agents and Agent Systems are steadily transitioning from small research projects to mainstream industrial and government markets. Multi-agent systems will live up to the expectations of implementing an openly-architected replacement of legacy stove-piped systems by providing flexible, customizable, user-friendly network-centric solutions. However, most of the currently available multi-agent systems do not address sufficiently the requirements and needs of information assurance. The problem is exacerbated in the DoD, where security requirements are stringent and agents are looked upon as unsafe. potentially harmful entities. To solve this problem in the context of large network-centric enterprise services (NCES) applications in the Navy (and DoD-wide), we propose a tool-based information assurance approach for complex multi-agent systems. By addressing IA as a first-class concern throughout the process of system engineering we enable the construction of systems that are automatically assured based on the certification the tools used to build and test those systems. Thus, the system's IA properties will emerge as a natural side-effect of engineering it in a rigorous manner, using the validated and certified Information Assured Intelligent Agents (IA2) tool suite. This suite of software tools, including analysis and transformation engines for software agents, is the core of the proposed technology.