一种基于内容流行度的内容中心网络缓存概率置换策略

发布时间：2018-09-14 06:52

【摘要】：内容中心网络是下一代互联网架构的研究热点,该架构通过分布式内容缓存机制实现以内容为核心的数据传输,改变了传统基于主机的网络通信模式。缓存置换策略是内容中心网络的核心研究问题之一,缓存置换策略的设计优劣直接制约内容中心网络的数据传输性能。该文针对内容的流行度分布特征,提出一种基于流行度的缓存概率置换策略,并针对L层级联的内容中心网络(CCN),推导了该策略下的网络分层缓存请求失败概率近似计算公式。该文策略根据请求数据块的流行度而选择数据块在缓存队列中的置换位置,尽可能平衡不同流行度内容在网络中的分布。数值计算结果表明:该策略适用于内容请求集中的网络应用,相比较传统的最近最少使用(LRU)策略,该文策略可以明显改善流行度较低的网络访问性能。
[Abstract]:Content-centric network is a hot topic in the next generation Internet architecture. This architecture implements content-centric data transmission through distributed content caching mechanism, which changes the traditional host-based network communication mode. Cache permutation strategy is one of the core research problems in content-centric network. The design of cache permutation policy directly restricts the data transmission performance of content-centric network. In this paper, a cache probability permutation strategy based on popularity is proposed, and an approximate formula for computing the failure probability of hierarchical cache requests is derived for the L-level content-centric network (CCN),) based on the characteristics of content popularity distribution. According to the popularity of the request data block, this strategy selects the replacement position of the data block in the cache queue, and balances the distribution of the different popularity content in the network as far as possible. Numerical results show that the proposed strategy is suitable for network applications in content-request set. Compared with the traditional (LRU) strategy, the proposed strategy can significantly improve the network access performance with low popularity.
【作者单位】：南京邮电大学通信与信息工程学院;江苏大学计算机科学与通信工程学院;南京邮电大学宽带无线通信与传感网技术教育部重点实验室;
【基金】：国家自然科学基金(60872018,60902015) 国家科技重大专项(2011ZX03005-004-03) 江苏高校优势学科建设工程资助课题
【分类号】：TP393.02

【相似文献】