Supporting the Requirements for Multilevel Secure and Real-Time Databases in Distributed Environments

Conflicts in database systems with both real-time and security requirements can sometimes be unresolvable. We attack this problem by allowing a database to have partial security in order to improve on real-time performance when necessary. By our definition, systems that are partially secure allow security violations between only certain levels. We present the ideas behind a specification language that allows database designers to specify important properties of their database at an appropriate level. In order to help the designers, we developed a tool that scans a database specification and finds all unresolvable conflicts. Once the conflicts are located, the tool takes the database designer through an interactive process to generate rules for the database to follow during execution when these conflicts arise. We briefly describe the BeeHive distributed database system, and discuss how our approach can fit into the BeeHive architecture.