Techniques Of Privacy Preserving Based On K-Anonymity
In location based services on mobile the k anonymity tries to preserve the identity of the sender by generating k-1 same requests, so that exact user can not be identified. Bettini propose a framework based on location based quasi identifier concept related with k-anonymity. Quasi identifier generate similar at least k-1 requests so that original requester’s identity is preserved. The work of location anonymizer is masking the location based information of the user thus preserving the privacy. It uses the query processor which can manage the anonymized queries. Gedik and Liu created a model for user to specify his or her minimum level of anonymity given as input to the engine which preserves anonymity. This is achieved by hiding user’s identity and spatio-temporal obfuscation of location information. Ghinita et al. proposed a decentralized architecture called as PRIVE and an algorithm to protect user’s identity. The algorithms works on the concept of k-anonymous areas through the Hilbert space-filling curve. Zhong and Hengartner proposed a distributed protocol for sender k-anonymity based on cryptographic mechanisms and secure multiparty computation.
The user interacts with multiple servers and a third party to determine if at least k people are in his or her area before communicating with the Location Based Services. The k-anonymity exhibits its disadvantages gradually, such as being easily attacked by continuous queries attacking algorithm, the larger the k value for higher security level lead to more cost of bandwidth and load of LBS server.
The K-anonymity is difficult to guarantee the user’s privacy for the continuous LBSs. For example, when the user issues LBS requests in continuous query points, and each point on the trajectory will form the cloaking region that satisfied the K-anonymity. However, the user trajectory can be easily disclosed under the following circumstances: (1) If the adversary links these cloaking regions, the user’s entire trajectory will be exposed. (2) If the adversary compares the users of each cloaking region at different query points, it can also recognize the real user. A DKM scheme in continuous LBSs, which introduces multiple anonymizers and combines the location selection mechanism with K-anonymity technology to improve the user’s trajectory privacy. The architecture of DKM scheme is shown (see Fig. 2), and the main components of the architecture are made up of three main entities: the user, the multiple anonymizers and the LBS servers.
At the user side, the solid line with an arrow indicates the user’s trajectory, the star on the trajectory represents the query point, the red dots and the green dots on the trajectory represent the predicted locations and the dummy locations respectively. We describe the three main entities and their interactions as follows: a) THE USER: The user is the devices with the capabilities of computation, memory and wireless communication, which acts as mobile devices with global positioning functionality (e. g. ,GPS). The user can obtain continuous LBSs from our system by issuing continuous query at different time points to different LBS servers through multiple anonymizers. At the same time, the user can look for (K −1) other users around his own through collaborative communication.
b) THE MULTIPLE ANONYMIZERS: The multiple anonymizers are placed between the user and the LBS servers, and their main function is to forward the user’s query requests and results in continuous LBSs. In our system, the multiple anonymizers have the K-anonymity function for the user’s location to ensure the user’s privacy in the LBS servers. At the same time, the user use the location selection mechanism to confuse the user’s real location to enhance the user’s privacy, and an anonymizer alone does not know the user’s trajectory, thus it can be the semi-trusted third party.
c) THE LBS SERVERS: The LBS server can be a service provider which has service databases to store and updates the service data and can provide kinds of data services for users. When the LBS server receives the location information and service queries from user, then searches for requested service data in the database, and replies the search results back to the user. In our system, there are different LBS servers to serve the user. The advantage of the DKM scheme is that user trajectory cannot be acquired from the LSP or a single anonymizer. In addition to offering more protection to the user’s trajectory privacy, it provides an effective approach for the anonymizer’s performance bottleneck and the single point of failure problem.
