USING COMPOSIBILITY TO IMPROVE SECURITY ANALYSIS TECHNIQUES

This proposal describes an approach for applying existing composibility techniques to analyze complex, multilevel secure (MLS) systems. Although there has been significant progress in developing composibility techniques, there is little experience with using these techniques. Thus, the set of problems to which the techniques are applicable and the amount of effort required to apply the techniques are unclear. The proposed project will examine the applicability of these techniques to computer security problems such as analyzing 1) distributed security enforcement, 2) assured pipelines for trusted subsystems, and 3) covert channels in distributed systems. For each of these potential problem areas, it is necessary to investigate the value of composibility both for synthesis and refinement. Instances in which composibility simplifies synthesis provide the basis for a library of reusable secure components. Instances in which composibility simplifies refinement suggest the use of modular decomposition to simplify security analysis. Based on deficiencies identified from this investigation, we will develop approaches for addressing the deficiencies. In summary, the goals of the proposed project are to 1) determine the problem areas to which composibility techniques are applicable and 2) address deficiencies in the prior work.