Time warp: How time affects privacy in LBSs

Abstract

Location Based Services (LBSs) introduce several privacy issues, the most relevant ones being: (i) how to anonymize a user; (ii) how to specify the level of anonymity; and, (iii) how to guarantee to a given user the same level of desired anonymity for all of his requests. Anonymizing the user within k potential users is a common solution to (i). A recent work [28] highlighted how specifying a practical value of k could be a difficult choice for the user, hence introducing a feeling based model: a user defines the desired level of anonymity specifying a given area (e.g. a shopping mall). The proposal sets the level of anonymity (ii) as the popularity of the area-popularity being measured via the entropy of the footprints of the visitors in that area. To keep the privacy level constant (iii), the proposal conceals the user requests always within an area of the same popularity-independently from the current user's position. The main contribution of this work is to highlight the importance of the time when providing privacy in LBSs. Further, we show how applying our considerations user privacy can be violated in the related model, but also in a relaxed one. We support our claim with both analysis and a practical counter-example. © 2010 Springer-Verlag.