Privacy Preserving Interceptor for Online Social Media Applications

Abstract

In the current scenario, online social network (OSN) plays a vital role in user’s day to day life. User’s store lots of their personal information in OSN, and they share their day to day experiences with their connections. There are enormous number of third party application (TPA) services allied with OSN to provide extended services to the users. The OSN manages the identity verification by authenticating the user and granting access to allied TPAs. The TPA requests permissions to access personal attributes about the user when accessed by the user for the first time. The personal attributes marked as required by the TPA has to be shared to avail the service. The privacy risk increases exponentially with the users TPA usage. The users not only leak their information to TPA but also end up unlocking a new type of threat from correlation with auxiliary information, through the data available from alternative sources. In this paper, we focus on reducing the level of sensitive data exposed to the external parties. The feasibility of providing such a service by restricting the data flow through access control policies is not feasible with the current All or Nothing approach. Hence, in this paper, we propose, the Privacy Preserving Interceptor (PPI) that acts as an interceptor between OSN and TPA to provide the required utility and yet preserves user’s privacy. PPI identifies the sensitive attributes shared by the user and transforms the original data into a less sensitive form that still meets the utility goals. Standard Differential Privacy in combination with other perturbation mechanism of replacing with random values is used in PPI. The users privacy remains more or less the same both before and after the data share to TPA and still meets the utility needs of the user.