Joint Optimization for Social Content Dissemination in Wireless Networks

Over the last decade, success of social networks has significantly reshaped how information disseminate among  users in terms of scale, intensity and speed. Much effort, both academic and industrial, has been dedicated to learning the patterns of content dissemination to further facilitate it.  However, as users become dominantly mobile in this process, little is done to consider the impact of the wireless environment, in particular, how to optimize the system, in the presence of capacity shortage, to better disseminate the contents. In this paper, we propose  a novel method, utilizing: 1) hybrid systems to handle active disseminating requests; and 2) predictions of dissemination dynamics from the social network applications. This method could mitigate the performance degradation for content dissemination due to wireless delivery delay. Results indicate that predictions of
dissemination at social layer, even coarse ones, benefit the social dissemination, by improving scheduling at wireless layer, which in turn improves the accuracy of the predictions with feedback.