The non-primitiveness of the simple-security property and its non-applicability to relational databases

Many works on secure databases state the simple-security-property as a postulate the necessity of which for achieving confidentiality seems to be out of question. We show that this property is not a postulate but a conclusion the validity of which relies an the assumptions made about the underlying secure environment. We then demonstrate that it contradicts the assumptions made by relational and logic-based databases. This result indicates that the semantic problems secure databases struggle which can be partially attributed to the accession of the simple-security-property