Inference attacks mean that a user infers (or tries to infer) the result of an unauthorized query execution using only authorized queries to the user. We say that a query q is secure against inference attacks by a user u if there exists no database instance for which u can infer the result of q. The security problem against inference attacks has been formalized under a model of object-oriented databases called method schemas. It is known that the technique of type inference is useful for deciding the security. However, the relationship of type inferability and decidability of the security has not been examined. This paper introduces a subclass of method schemas, called linear schemas, and presents the following results. First, type inference of linear queries is possible under linear schemas. Next, the security of type-inferable queries is undecidable under linear schemas. Moreover, type inference is impossible for queries whose security is decidable under linear schemas. These results imply that type inferability and decidability of the security problem are incomparable. © Springer-Verlag 2004.
CITATION STYLE
Ishihara, Y., Shimakawa, Y., & Fujiwara, T. (2004). Type inferability and decidability of the security problem against inference attacks on object-oriented databases. Lecture Notes in Computer Science (Including Subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), 3269, 145–157. https://doi.org/10.1007/978-3-540-30191-2_12
Mendeley helps you to discover research relevant for your work.