Instant Anonymization

Anonymization-based privacy protection ensures that data cannot be traced back to individuals. Researchers working in this area have proposed a wide variety of anonymization algorithms, many of which require a considerable number of database accesses. This is a problem of efficiency, especially when the released data is subject to visualization or when the algorithm needs to be run many times to get an acceptable ratio of privacy/utility. In this paper, we present two instant anonymization algorithms for the privacy metrics k-anonymity and l-diversity. Proposed algorithms minimize the number of data accesses by utilizing the summary structure already maintained by the database management system for query selectivity. Experiments on real data sets show that in most cases our algorithm produces an optimal anonymization, and it requires a single scan of data as opposed to hundreds of scans required by the state-of-the-art algorithms.