A brief overview of basic inference attacks and protection controls for statistical databases
Olga Dzięgielewska, Bolesław Szafrański · Computer Science and Mathematical Modelling · 2016
With cyber-attacks on the dramatic rise in the recent years, the number of entities which realize the necessity of protecting their IT assets increases. Individuals are more aware of the potential threats and demand high level of security from the business entities having access to their personal and private data. Such entities have legal obligations to satisfy the confidentiality when processing sensitive data, but many fails to do so. Keeping the statistical data private is a challenge as the approach to the security breaches slightly differs from the classical understanding of data disclosure attacks. The statistical disclosure can be achieved using inference attacks on the not-effectively protected assets. Such attacks do not target the database access itself, i.e. are performed from a perspective of an internal user, but the statistical interface used to retrieve the statistical data from the database records. This paper sums up basic types of inference attacks classifying them in the CVSS standard and provides a series of fundamental countermeasures which can be undertaken to mitigate the risk of performing successful attack.