Context-dependent Security Enforcement of Statistical Databases

摘要

A statistical database system is a database system that contains information about individuals (companies, organisations) which enable its authorized users to retrieve aggregate statistics such as sample total, mean and count. The security problem for a statistical database is to limit the use of the database so that only statistical information is available to an authorized statistical user and no sequence of queries is sufficient to infer protected information about any individual. If, however, as a result of a statistical query, individual confidential information is obtained by a statistical user, then the database is said to be compromised. In order to prevent compromise, we use a knowledge based system approach. A statistical user can pose only statistical queries to the database. The knowledge based system will infer a query result by answering questions from the statistical user in such a way that the individual information is kept confidential. Earlier models modelled SDB compromise using two kinds of knowledge [7]. In this paper we enhance this model by including another important knowledge, namely, legal knowledge, and we describe an implementation of a knowledge base that supports the decision making for the protection of privacy in statistical databases, while taking into account not only the working and (known) supplementary knowledge but also legal knowledge.