Maintaining surrogate data for query acceleration in multilevel secure database systems

Abstract

Various models have been developed for classified data management in a multilevel secure database system. While concurrency control policies have been the focus of these models, the PRISM model was developed to solve the query delay problem in such systems. In order to accelerate multilevel queries, the PRISM model maintains surrogate data in different relations, as opposed to the actual data values, that result in higher record density in a page, thus requiring less number of page I/Os. In order to establish mapping between the surrogate values and the actual data values some auxiliary data structures are maintained. This paper presents the cost of maintaining such surrogate values and corresponding data structures in the model. The result is compared with the maintenance cost of the relations in the Sea View model that has been used as the base model for the PRISM model.