The area of transaction processing in multilevel secure (MLS) database management systems (DBMSs) has substantially progressed in the past few years. Despite this activity, research in this area has been limited to traditional transaction concepts. Moreover, even for traditional transaction models, existing solutions for secure transaction processing are not completely satisfactory. This research (1) investigates the impact of multilevel security constraints on transaction processing in advanced database applications, and (2) examines whether the properties of the advanced transaction models, in particular nested transaction models, can be utilized to solve the secure concurrency control problems in traditional databases. It develops multilevel secure workflow and multilevel secure long duration transaction models, and the necessary extensions to the nested model required by MLS DBMSs. The results of this project enable incorporating multilevel security in advanced database applications such as process modeling, engineering design and long-running activities, and provide efficient secure transaction processing protocols for conventional database applications. Besides MLS databases, the results of this work are directly useful in other non-secure database environments such as hierarchically decomposed databases. The education plan includes development of three new courses at undergraduate and graduate level and student involvement in this research.
