面向ICN切片的视频缓存与传输服务

Integrated ICN Slice and Video Caching as a Transmission Service

随着视频流量爆炸式地增长,为了降低密集区域用户的视频检索时延和网络负载,利用网络功能虚拟化技术提出了一种将5G切片与信息中心网络(Information Center Network,ICN)相结合的内容分发服务系统集成的架构,并改进基于ICN的缓存策略。该缓存策略依据特定区域用户一个月与一周内的历史访问记录构建视频主题的兴趣函数,并结合互联网视频流行度构建ICN切片的有效缓存函数,为密集区域用户针对性地缓存视频内容。仿真结果表明,所提算法能够有效降低视频检索时延和提高视频缓存的命中率,ICN切片的视频检索时延最多低于IP传输系统时延11.1%;相比虚拟类聚类的最流行缓存策略最多可提升16%的缓存命中率,最多可降低20.9%的检索时延;相比移动边缘计算分组的协作缓存策略最多可降低15.5%时延并提高19.3%的缓存命中率。

With the explosive growth of video traffic,in order to reduce the delay of video retrieval and network load of users in dense areas,a content distribution service system integrating 5G slice with Information Center Network(ICN)is proposed by using Network Function Virtualization(NFV),and the cache strategy based on ICN is improved.The cache strategy builds the interest function of video theme based on the historical access records of users in a specific region for a month and a week,and constructs an effective cache function of ICN slice based on the video popularity of the Internet,so as to cache video content targeted for users in dense regions.Simulation results show that the algorithm can effectively reduce the delay of video retrieval and improve the hit rate of video cache.The video retrieval delay of ICN slice is at most 11.1%lower than that of IP transmission system.Compared with Cluster Most Popular Caching(CMCP),the cache hit rate can be increased by 16%and the retrieval delay of 20.9%can be reduced at most.Compared with Mobile Edge Computing(MEC)group cooperative cache strategy,the proposed strategy can reduce the delay by up to 15.5%and increase the cache hit rate by 19.3%.