The load on mobile networks increased over the past years and is predicted to further grow rapidly. Mobile network operators are facing new challenges to deliver mobile data in a user-satisfying way. This situation can be improved, e.g., by offloading traffic to WiFi networks or by shifting traffic to times where good network conditions can be leveraged, e.g. non-peak hours. One way to offload content is speculative prefetching. By knowing which contents will be requested in the future allows to prefetch it in advance. To this end, the users’ content consumption has to be predicted. Furthermore, predicting properties of the smartphone’s mobile connectivity, e.g., strong signal, allows a better user experience as well as an optimized network resource allocation. The reason for this is that the packet loss, retransmission and congestions can be avoided. To this end, the Mobile Social Prefetcher app aims at relieving the mobile network in a two-fold way. On the one hand, prefetching of promising videos from the users Online Social Networks is performed if WiFi is available. Videos posted on the user’s OSN have been shown likely to be watched. On the other hand, if no prefetching opportunity can be used, a specific network optimized streaming is performed. This is provided by a video player which, both, reduces the load for the network operator
and decreases stalling events of video playbacks. The proposed approach prefetches videos from video posts on the user’s Facebook feed. Furthermore, the app considers
current and previously observed cellular network information of the smartphone to optimize the mobile data throughput. This way, both, the operators’ and the users’ needs are reflected by the approach demonstrated.