Slicing who slices: Anonymization quality evaluation on deployment, privacy, and utility in mix-zones

In the flowering of ubiquitous computing, technologies like the Internet of Things and the Internet of Vehicles have contributed to connecting objects and sharing location services in broad environments like smart cities bringing many benefits to citizens. However, these services yield massive and unrestricted mobility data of citizens that pose privacy concerns, among them recovering the identity of people with linking attacks. Although several privacy mechanisms have been proposed to solve anonymization problems, there are few studies about their behavior and analysis of the data quality anonymized by these techniques. This paper presents an anonymization quality framework for mix-zones enabling characterizing and evaluating the impacts of anonymization over time and space in mobility data. We conducted experiments with a cab mobility dataset and two positioning algorithms to explore one of the potentialities of the anonymization quality: elect mix-zones that do not consider the traffic but its operating requirements too. The results showed that the anonymization quality enabled the selection of mix-zones that yield data anonymization considering the quality, privacy, and utility analysis. This study is unique because it analyzes mix-zone coverage and quality metrics to observe the anonymization quality not found in the literature.