Secure Computer Systems: Extensions to the Bell-La Padula Model

摘要

This report provides a summary of initial results of a project investigating solutions to problems in flowing valued information among coalition partners. The research objectives of the Flowing Valued Information project include: (1) improving our capability to enable automated understanding of command intent and (2) improving our capability to provide automated support of a command decision to share information. Initial investigations have indicated a need to extend the mathematical foundations provided by D. Elliott Bell and Leonard J. La Padula which applied early system theory to enable building formal systems for proving security results for distributed computing systems. Our extensions are in two areas: (1) we discuss application of current system theory results in modeling compositions of continuous and discrete systems, and (2) we discuss mathematical foundations for adding support for a commander's decision to share information. The motivation for the extensions is grounded in two continuing shortfalls in science and technology available for decision support: (1) the inability of current system models to predict future state of complex systems and (2) the continued difficulty in enabling automated support for a commander's decision to share information in order to meet mission requirements. We believe that useful extensions are achievable for building more accurate models of complex system interactions for small unit operations since general system theory has advanced since the work of Bell and La Padula and we believe that explicit extensions for sharing information are needed (and possible) for information which needs to be shared while simultaneously protecting information which must remain protected.