Modeling Trusted Processing Architectures for Mandatory Access Control

Abstract

This paper introduces a trusted architecture graph (TAG) model, which can be used for modeling the semantics of trusted architectures designed to enforce mandatory access control. The TAG permits the modeling of various types of trusted functions, storage functions and processing functions and their interconnection through various types of links. The value of the TAG and the associated TAG notation is that they provide a uniform way of representing different trusted architectures that may be described either informally in a natural ll.lnguage, or formally (but voluminously) described in design documents or programming code. By providing a concise yet expressive description of the architecture, the various features of one architecture can be readily compared with another's. This paper provides some examples that illustrate how various trusted database management system architectures can be formulated in the TAG notation such that their significant differences can be readily observed.