Linear queries in statistical databases
M. D. Schwartz, Dorothy E. Denning, Peter J. Denning · ACM Transactions on Database Systems · 1979
A database is compromised if a user can determine the data elements associated with keys which he did not know previously. If it is possible, compromise can be achieved by posing a finite set of queries over sets of data elements and employing initial information to solve the resulting system of equations. Assuming the allowable queries are linear, that is, weighted sums of data elements, we show how compromise can be achieved and we characterize the maximal initial information permitted of a user in a secure system. When compromise is possible, the initial information and the number of queries required to achieve it is surprisingly small.
Cited by 38 works. doi:10.1145/320071.320073 Open-access PDF
Authors
References
- Secure databases
- The tracker
- A model of statistical database their security
- Even Data Bases That Lie Can Be Compromised
- Identification and Retrieval of Personal Records from a Statistical Data Bank
- Security in statistical databases for queries with small counts
- Confidentiality of Statistical Records: A Threat-Monitoring Scheme for On Line Dialogue
- Insuring individual's privacy from statistical data base users
- Inference from statistical data bases.
- Securing Data Bases under Linear Queries.