A Privacy-Preserved Probabilistic Routing Index Model for Decentralised Online Social Networks

Abstract

Despite the tremendous success of online social networks (OSNs), centrally controlled OSNs have inherent issues related to lack of user privacy and single point of failure. These limitations have motivated the research community to shift the computing paradigm from a centralised architecture to decentralised alternatives. Existing research works mainly focused on the routing mechanisms using social information in decentralised OSNs, without considering the user’s privacy. This paper proposes a self-organised decentralised architecture (SDA) that leverages privacy-preserved routing methods to facilitate query routing in decentralised social networks. This architecture encompasses a hash-based profiling model to characterise semantic features of the user’s content with low dimensionality and privacy-aware mechanisms to organise similarity users into semantic communities. Furthermore, a probabilistic routing method is proposed to support efficient information dissemination and service discovery. The correctness and efficiency of our proposed approach are evaluated through simulations on real-world datasets. The experimental results demonstrated that our approach achieved a better topological structure with high routing efficiency.

Bo Yuan

Lecturer in (Assistant Professor) Computer Science

My research interests include Data Science, Artificial Intelligence, Machine Learning, Internet of Things, Distributed Computing, Edge Computing