With the mobile data traffic grows rapidly and the video data has a high proportion, the backhaul link faces great pressure. The conventional centralized architecture has been far from enough to satisfy the user demands. For reducing user response latency and easing backhaul stress, the streaming media contents should be proactively stored in the edge of network and new contents caching model needs to be presented. In this work, a video content collaborative caching strategy in the cloud-edge cooperative environment is proposed. In this strategy, first, the k-means algorithm is used to cluster the edge servers. Then the latency and caching cost gain bring by caching the content on the edge servers in the cluster and the collaborative cache domain are analyzed to establish the content caching problem. Further, the marginal gain is calculated by analyzing the latency and caching cost gain. Finally, in order to solve the content caching problem, this paper proposed a marginal gain based content caching algorithm. Experimental results prove the effectiveness of the proposed algorithm.

随着移动数据流量快速增长，视频数据占比高，回传链路面临巨大压力。传统的集中式架构已经远远不能满足用户的需求。为了减少用户响应时延，缓解回传压力，流媒体内容应主动存储在网络边缘，需要提出新的内容缓存模型。在这项工作中，提出了一种云边缘协同环境下的视频内容协同缓存策略。在该策略中，首先使用 k-means 算法对边缘服务器进行聚类。然后分析在集群边缘服务器和协同缓存域上缓存内容带来的延迟和缓存成本增益，建立内容缓存问题。更多，边际收益是通过分析延迟和缓存成本收益来计算的。最后，为了解决内容缓存问题，本文提出了一种基于边际增益的内容缓存算法。实验结果证明了所提出算法的有效性。