内容分发网络下的用户原创视频热度研究

视频的热度是指用户对视频的关注程度.主要对大规模内容分发网络环境下的在线共享网站原创视频的热度进行分析.提出了综合视频点击数、评论数、评分等多元指标对视频热度进行统计刻画,分析各个指标的分布模型以及指标之间的相互关系,并对视频热度随时间的变化趋势进行了详细分析.通过对实际的爬虫数据统计分析,发现热度指标的分布具有典型的无标度特征,随着视频上传时间的增长,热度越来越趋于稳定.上述发现对于视频内容提供商和互联网运营商在服务器部署、内容缓存、容量规划和流量优化等方面具有重要的意义.

Characterizing Popularity of User-generated Videos CDNs

The popularity of user generated videos means how users concern about the videos. This paper presents a characterization of the popularity of user generated content videos in large-scale content delivery network. We study the view counts, comments, ratings as indicators to depict the popularity of video. We illustrate the probability distribution and interrelation of the probability indicators. We find that they have the scale-free characterization. We also present the temporal trend of popularity along with the upload time. We find that the popularity gradually level off with the increase of upload time. The experimental results have great significance in server deployment, content caching, capacity planning and traffic optimization for video content providers and network operators.