Rendezvous Avoidance File Caching for Mobile Social Networks

File caching and replications for mobile social networks (MSNs) play a critical role in accessible and faster file sharing for mobile cloud, where copies of an original file are duplicated and stored in caching nodes with high network centrality. However, the existing solutions are not disease- and infection-resilient, i.e., the caching performance in terms of file availability and query delivery rates will decrease in the case of a disease outbreak. In such contexts, the nodes with high centrality as well as their direct neighbors in a social network are likely to go offline due to some reasons, such as disease infection. As a result, caching nodes with the same cached files can be accidentally removed from an MSN at the same time. To tackle this issue, in this article, we first introduce the concept of rendezvous avoidance, in which not only the connectivity among nodes but also the distances among caching nodes are incorporated into the centrality design. Then, we propose the rendezvous avoidance file caching (RAFC) algorithm for MSNs, where some nodes may accidentally go offline. In RAFC, the scores of nodes are iteratively computed by the proposed rendezvous avoidance centrality, and then, caching nodes are selected. The performance of our proposed scheme is evaluated by simulations using real mobility traces, and the simulation results demonstrate that it achieves its design goals.