A Method for Model-driven Information Flow Security

We present a method for software development in which information flow se-curity is taken into consideration from start to finish. Initially, the user of the method (i.e., a software developer) specifies the system architecture and selects a set of security requirements (in the form of secure information flow properties) that the system must adhere to. The user then specifies each component of the system architecture using UML inspired state machines, and refines/transforms these (abstract) state machines into concrete state machines. It is shown that if the abstract specification adheres to the security requirements, then so does the concrete one provided that certain conditions are satisfied.