基于区域划分的混合式NDN缓存策略  

Hybrid NDN caching strategy based on region division

作　　者：宋彩风 李翀[1] 刘学敏[1] Song Caifeng;Li Chong;Liu Xuemin(Computer Network Information Center,Chinese Academy of Sciences,Beijing 100190,China;University of Chinese Academy of Sciences,Beijing 100049,China)

机构地区：[1]中国科学院计算机网络信息中心,北京100190 [2]中国科学院大学,北京100049

出　　处：《计算机应用研究》2019年第7期2124-2129,共6页Application Research of Computers

摘　　要：针对NDN(命名数据网络)中确定性缓存和概率性缓存进行研究,提出一种确定性缓存和概率性缓存相结合的混合式NDN缓存策略(HDP)。基于区域划分的思想,在网络边缘采用基于热度的确定性缓存策略,在网络核心采用基于缓存收益和内容热度的概率性缓存策略,从而将两种缓存策略的优势相结合,进一步提高NDN缓存性能。仿真实验表明该策略与现有NDN缓存方法相比,能有效提高缓存服务率和命中率,并有助于降低内容访问延迟,改善用户体验。Considering the advantages and disadvantages of both deterministic caching strategies and probabilistic caching strategies in named-data networking,this paper proposed a hybrid caching strategy (named HDP).Based on the idea of region division,the hybrid NDN caching strategy used the popularity-based deterministic caching strategy at the edge of the network,and used the probabilistic caching strategy based on cache benefit and content popularity at the core area.Combining the advantages of both cache strategies,it enhanced the performance of NDN caching.Simulation results show that compared with current NDN methods this strategy is able to increase cache service ratio and hit ratio and reduce content access delay to improve the user experience.

关 键 词：命名数据网络 缓存技术 确定性缓存 概率性缓存 混合式缓存 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]