An Algorithm on Video Caching Design in Content Delivery Network
首发时间:2017-06-16
Abstract:In recent years, multimedia traffic such as video on demand shows explosive growth on the Internet. Network operators are considering content delivery network as the main strategy to keep up with the increasing demand for video requests with the minimum possible delay. Layered video encoding strategies such as Scalable Video Coding enable different caches to store different layers of a video content, allowing successive improvement of the video stream quality. However, the caching placement of video encoding layers that minimize the average delay is NP-hard. We reduce the complexity of this problem to pseudopolynomial time by using a connection with multiple-choice knapsack problem. Numerical results show up to 25% reduction in delay over existing video caching strategies.
keywords: Network Design, Video Encoding, Cache Design
点击查看论文中文信息
一种CDN视频缓存设计的优化算法
摘要:现代信息科技的发展带来的是多媒体内容流量的爆炸性增长,其中以视频点播业务为代表。为了尽量降低这种现状所带来的网络延迟问题,网络运营商通常以构建内容分发网络(CDN)作为网络加速方案。其中分层视频编码策略作为CDN配套方案的一种新策略,能够使得视频不同层面的媒体信息储存通过不同层面的缓存系统各自实现,提高了视频流的编解码效率。但是,关于视频解码系统中如何配置缓存结构的层级,使得平均传输延迟最小化,这是一个NP困难问题(NP-hard problem),无法直接求出最优解。本文将此问题与算法中的多选择背包问题结合,试图将求其有损地简化为一个多项式时间内可求解的问题。数据结果显示,与常见的缓存策略比较,此算法平均缩短了 25% 的延迟时间。
论文图表:
引用
No.4735355119191514****
同行评议
共计0人参与
勘误表
一种CDN视频缓存设计的优化算法
评论
全部评论0/1000