A Logical Formalization of Integrity Policies for Database Management Systems

This paper presents a model to specify integrity policies for database management systems. This model makes it possible to (1) assign an integrity level to each user -this integrity level depends on the data this agent is authorized to update, (2) define updating permissions and prohibitions associated with each user -in particular we show that permission and prohibition to update may be independent from the user’s integrity level, (3) define a policy to manage how integrity evolves in time. Our model is compared with classical approach, such as Biba and Clark-Wilson. In particular, we do not follow Biba: in our model, a subject may be authorized to update data even if its integrity level is not higher than or equal to the integrity level of the data.