Relationship-based information sharing in cloud-based decentralized social networks.

ABSTRACTCommercial OSNs have started to provide users with the ability to set their privacy settings for a more controlled information sharing. However, these settings do not prevent the social network manager to perform marketing research on user personal data, aiming, as example, at offering a personalized advertising to users. To cope with these requirements Decentralized Social Networks (DSNs) are emerged as a possible solution for moving users' personal data out from OSN realms. Unfortunately, it has been shown that DSNs have some limitations, in terms of usability and social features they offer. To overcome this problem, in this paper we extend the DSN framework so that users' data (e.g., resources and relationships) are securely stored in a public cloud data storage and shared according to relationship-based rules defined by owners, by at the same time supporting a privacy-preserving path finding. To this end, we make use of encryption techniques and we devise a new collaborative anonymization process. In the paper, besides presenting all the components of our framework, we analyze its security and present experiments showing the feasibility of the developed techniques.