Nowadays, smartphones are used to download files such as movies, music, etc. These files present some flexibility in delivery time; we exploit this feature to reduce network overload peaks. If the network is overloaded, downloads are postponed, thus reducing the traffic load during rush hours. We propose an architecture that enables the network to determine the best time to trigger a download that is queued. This architecture is based on a specific server in the network and a mobile application. This contribution describes the proposed architecture, presents the developments that have been made, and analyzes the performance of this novel concept.