命名数据网络中基于数据请求节点的就近缓存算法

Consumer-based proximity caching algorithm for named data networking

为提高命名数据网络(named data networking,NDN)中数据存储节点的缓存效率和存储空间利用率,根据就近缓存思想,提出一种在数据请求节点中根据特定内容兴趣而区分缓存的数据缓存算法。该算法结合缓存节点地理位置差异性和缓存数据内容热度差异性,将热度高的数据内容优先缓存在数据请求节点周围,并根据数据被请求的频次动态设置缓存时间,使热度高的数据内容尽可能长时间地缓存在存储节点中,增大了数据就近响应概率;且节点中被替换的数据内容没有直接删除,而是向上转发并指示上游节点缓存,增加了缓存数据内容的多样性。仿真结果表明,所提算法能有效提高数据搜索命中率,降低网络请求时延以及路由传输跳数。

To improve data caching efficiency and storage space utilization in named data networking( NDN) storage nodes,based on proximity caching idea,a differentiated data caching algorithm is proposed in this paper to provide differentiated caching service according to specific content for consumer nodes. The proposed algorithm combines the location difference of the cache node and the cache data content popularity difference,so that it can put the popular data content around the consumer nodes first,and the cache time is set dynamically based on the requested frequency of data. Then the data content of the high popularity can be cached in the storage node for as long as possible,thus the data proximity response probability is enlarged; instead of deleting the data in the node directly,it forwards and directs the upstream node cache,increasing the diversity of cached data content. Simulation results show the proposed algorithm can effectively achieve high cache hit ratio,reduce data response delay and the number of hops.