边缘网络环境下基于流量分类的SDN-NDN缓存策略

SDN-NDN Caching Strategy based on Traffic Classification in Edge Network

在边缘网络环境中使用命名数据网络(Named Data Network,NDN)解决单点失效、通信高时延等问题是一种新的研究趋势,但是传统的NDN缓存策略所造成的缓存冗余高及缓存策略单一等问题始终没有得到完善的解决,因此提出了一种基于流量分类的SDN-NDN缓存策略。通过对边缘网络环境中的业务流量进行分类处理,优化SDN-NDN的转发模型和传统的NDN缓存策略。首先使用与流行度相结合的动态缓存方法提升了网络缓存的全局命中率,其次使用基于流量分类的缓存替换策略降低了边缘网络中的缓存冗余,最后结合流量分类进行关联性缓存提升了网络中缓存方法的多样性。仿真实验表明,该缓存策略具有可行性,且与其他NDN缓存策略相比,在全局命中率和缓存占用率方面更具优势。

It is a new research trend to use Named Data Network(NDN)in edge network environment to solve problems such as single point failure and high communication delay.However,problems such as high cache redundancy and single cache strategy caused by traditional NDN cache strategy have not been solved completely.Therefore,a SDN-NDN caching strategy based on traffic classification is proposed to solve the above problems.By classifying the traffic in the edge network environment,the forwarding model of SDNNDN is optimized while the traditional NDN caching strategy is optimized.Firstly,the dynamic cache method combined with popularity improves the global hit ratio of network cache.Secondly,the cache replacement strategy based on traffic classification reduces the cache redundancy in edge network.Finally,the correlation cache combined with traffic classification improves the diversity of cache methods in network.This caching strategy is feasible through simulation experiments,and compared with other NDN caching strategies,it has more advantages in global hit ratio and cache occupancy rate.