K-source: multiple source selection for traffic offloading in mobile social networks

The constantly increasing mobile data traffic imposes heavy pressure on the current cellular networks. Traffic offloading which uses complementary network communication technologies to deliver data traffic is a possible solution for this problem. And the efficiency of such opportunistic traffic offloading is highly determined by the selection of initial sources. As an effort towards this issue, we propose a multiple sources selection method to find both the optimal number of initial sources and who are the sources in traffic offloading applications in mobile social networks. Our designed cost function consider both the diffusion time and the cost of data transmission via the cellular networks. Moreover, to accurately estimate the diffusion time, we calculate the diffusion probabilities differently in the dense contact network scenario and the sparse contact network scenario. The parameters in different schemes are analyzed and the proposed scheme is evaluated by extensive simulations.