The prosperity of location-based services (LBSs) makes more and more people pay close attention to personal privacy. In order to preserve users privacy, several schemes utilized a trusted third party (TTP) to obfuscate users, but these schemes were suspected as the TTP may become the single point of failure or service performance bottleneck. To alleviate the suspicion, schemes with collaborative users to achieve [Formula: see text]-anonymity were proposed. In these schemes, users equipped with short-range communication devices could communicate with adjacent users to establish an anonymous group. With this group, the user can obfuscate and hide herself behind at least [Formula: see text] other users. However, these schemes are usually more efficient in snapshot services than continuous ones. To cope with the inadequacy, with the help of caching in mobile devices, we propose a query information blocks random exchange and results caching scheme (short for CaQBE). In this scheme, a particular user is hidden behind collaborative users in snapshot service, and then the caches further preserve the privacy in continuous service. In case of the active adversary launching the query correlation attack and the passive adversary launching the impersonation attack, a random collaborative user selection and a random block exchange algorithm are also utilized. Then based on the feature of entropy, a metric to measure the privacy of the user against attacks from the active and passive adversaries is proposed. Finally, security analysis and experimental comparison with other similar schemes further verify the optimal of our scheme in effectiveness of preservation and efficiency of performance.
A generalized correlation attack on a class of stream ciphers based on the Levenshtein distance
