Community-Centric Brokerage-Aware Access Control for Online Social Networks

Abstract

Widespread use and global outreach of Online Social Networks (OSNs) has made it indispensable for everyone to be a part of this global community. Users of OSNs want to share and disseminate some information to the general public, on the other hand, they also want to keep certain information hidden or restricted to a few people. An OSN allows users to join different groups based on their interests, however, some irrelevant people may also join these groups. This brings forth the challenge of privacy to OSNs. Therefore, an effective access control system is highly desirable that can manage the access control of incoming and outgoing users of a group in OSNs. This study deals with privacy issues of online communities and aims to support access control in community-based OSNs. We propose a Community-Centric Brokerage-Aware Access Control (CBAC) model that utilizes important concepts from Social Network Analysis (SNA) which are brokerage, one-to-many relationships, temporary relationships, as well as from emerging access control models such as attribute-based and trust-based access control along with decentralized policies. User attributes can play a vital role in detecting communities in a graph. For community detection, we propose Attribute-Based Community Detection (ABCD) algorithm and compare its results with Louvain, Newman’s Eigenvector, and Clauset algorithms. The results of the ABCD algorithm outperform other methods in terms of modularity and the number of communities for large and dense networks. The contribution of this study is to propose a CBAC model, its mathematical modeling and the ABCD algorithm for community detection. The performance of the CBAC model was evaluated in terms of grants, denials, and response time. The performance of the proposed the CBAC model is compared with the Role-based Access Control (RBAC) model and the Team-based Access Control (TMAC) model. The CBAC model offers the best data sharing capabilities and provides comparatively reasonable response time.