Information diffusion in heterogeneous networks: The configuration model approach

In technological or social networks, diffusion processes (e.g. information dissemination, rumour/virus spreading) strongly depend on the structure of the network. In this paper, we focus on epidemic processes over one such class of networks, Opportunistic Networks, where mobile nodes within range can communicate with each other directly. As the node degree distribution is a salient property for process dynamics on complex networks, we use the well known Configuration Model, that captures generic degree distributions, for modeling and analysis. We also assume that information spreading between two neighboring nodes can only occur during random contact times. Using this model, we proceed to derive closed-form approximative formulas for the information spreading delay that only require the first and second moments of the node degree distribution. Despite the simplicity of our model, simulations based on both synthetic and real traces suggest a considerable accuracy for a large range of heterogeneous contact networks arising in this context, validating its usefulness for performance prediction.