摘要

The explosive increase of the mobile video service has imposed a challenge, big data, for current cellular networks. D2D communication, as a key technology for the fifth mobile communication system (5G), provides a powerful platform for the large-scale video transmission. However, D2D communication also suffers from immense obstacles due to the limited storage capacity, the discrepant computational ability, the dynamic communication environment, the random network establishment, and the diverse services of the large-scale video applications. In this work, we propose a systematic solution for the large-scale video transmission over D2D communication. At first, through studying the relationship between the D2D communication service qualities and the video coding, a hybrid video coding is designed with flexibility, robustness, and simplicity. Subsequently, to resolve the fundamental contradiction between the limited storage capacities and the large-scale video contents, an efficient content updating and delivery strategy is proposed based on an appropriate prediction of the video popularity, priority, and requirement. In addition, a cooperative video scheduling scheme is developed to realize the dynamic optimization adaptation between the video streams and network resources.