Privacy And Security In Location-Based Social Networks

Nowadays, all kinds of social networks are becoming necessary part of people’s daily life. The broad masses use the social networks to establish instant messaging, find new friends, participate in the activity, discuss the sharing hobby, and post the daily life, etc. At the same time, there are lots of attacks against social networks’users and various methods are used to invading high valued user accounts. Moreover, most users don’t pay much attention to the potential threats of their social network activities and public information, so there exists huge hidden danger in social networks. Especially location information and the other user public information that come along with it, all can easily accessed by adversary and then leveraged to locate the user, profile the user, re-identify the user, and track the user’s moving trajectory, etc.Location information is always the focus of adversary in the location-based social discovery. Traditional social networks report the user’s location with exact meters, which is easily attacked by the Euclidean geometry algorithm based trilateration method. To thwart the trilateration attack, modern social networks adopt the concentric band-based approach when reporting distances. Thus, the trilateration is no longer useful. The revised trilateration uses three concentric rings to location a user in a rather small area. But this method is less accurate than the original one. And then there is partition algorithm, which is more accurate than the revised trilateration method but it is still not accurate enough.This paper gives thorough analysis on the privacy and security problems that the LBSD (Location-Based Social Discovery) services in LBSNs (Location-Based Social Networks) bring. The main work of the research includes:(1) Compared with traditional LBSNs locating algorithms, this paper studies the broadly used concentric band-based distances in modern LBSNs and the locating algorithms against it specifically, then summarizes the strengths and weaknesses of different algorithms and then comes up with a number theory based locating algorithm. The algorithm can efficiently pinpoint any users in LBSNs with coarse band-based distances. The accuracy of the algorithm can reach 1 meter theoretically. Then it take an example of WeChat to show the usability and accuracy.(2) This paper also investigates and analyzes the public user information that exposed together with the location information when users using the LBSD services. Then we design a system that captures user public information. The system can automatically crawl public information such as gender and username that comes together with location information. Then the system classifies the username into different category according to their characteristics, and quantifies the relationship between anonymity, gender and user behavior pattern. We further show that the LBSNs users not only under the potential danger of location expose but also behavior pattern.(3) According to the existing security problems of LBSNs, this paper extracts the specific attributes that LBSNs have, and then gives the detailed quantification criteria. At the same time, this paper summarizes the security and privacy requirements of LBSNs, integrates attribute-based encryption schemes, and proposes a proper protection scheme that suits LBSNs.