Preserving user data privacy while providing utility at the same time requires good system and protocol design and implementation in practice. This paper targets the anonymization of user data from distributed database systems belonging to different entities. The authors “propose a simple, efficient, and secure distributed protocol for the specific statistical disclosure control (SDC) method of rank shuffling.” Microaggregation is one of the most common methods used to obtain k-anonymity; it has provided good quality results such as the maximum distance to average vector (MDAV) algorithm; variable-size MDAV (V-MDAV), which is the improvement version of MDAV; and the centroid-based fixed-size (CBFS) algorithm. However, designing efficient distributed microaggregation protocols is a hard task. Rank swapping, another kind of SDC method, is simple and easy to use, but weak in terms of privacy against re-identification attacks. Also, through the experimental study of the protocol, the authors prove that their protocol “provides either more security or more efficiency” than other distributed versions of SDC methods.
The privacy problem of cooperation among different entities becomes more and more common and serious as many entities try to share information to achieve good utility and productivity. By investigating the existing popular SDC methods, the authors have found that many of them are very complicated due to inefficient cryptographic sub-protocols and highly depend on the number of bits of data. They implement a new distributed protocol using rank shuffling as the inherent SDC method.
The experiments show that the quality of the protocol is better than other protocols in terms of information loss (IL) and a probabilistic variation of IL (PIL). For efficiency, it takes less time for key generation and consumes less computing resources. Also, the authors conclude that collective strategy is better than individual strategy. Overall, the paper presents convincing and promising results for this specific problem.