Dispersing Social Content in Mobile Crowd through Opportunistic Contacts

Crowdsourced content sharing has become a fast-growing activity in today's online social networks, where contents of interest are created by diverse source users and conveyed over the network as friends view and reshare. The rapid and boundless propagation in a mobile crowd however often creates severe bottlenecks on the server side and incurs significant energy and monetary costs on the mobile side, particularly with the still expensive 3G/4G cellular connections. This paper presents SoCrowd, a novel framework for large-scale content sharing in a mobile crowd by exploiting contacts, i.e., users happen to move close with such short range low power communications as WiFi and bluetooth being enabled. We formulate the scheduling problem for social content propagation in a mobile crowd with contacts, and present optimal dynamic programming solution, which further motivates a series of practical heuristics. The effectiveness of SoCrowd has been demonstrated by extensive simulations driven by realworld traces and datasets.