一种利用K均值算法的内容中心网络缓存机制

Caching Mechanism Utilizing K-means Algorithm in Content-centric Networks

互联网的应用方式正从以面向主机的点对点通信为主转向以海量内容获取为主.为适应这一转变,研究界提出了以内容为中心(Content-Centric Networks,CCN)的新型网络架构.网络缓存作为内容中心网络的重要元素,能够有效减少网络带宽和传输时延.为了降低内容中心网络的缓存内容冗余度和提高缓存内容命中率,提出了一种利用K均值算法的缓存机制KCache(K-means algorithm utilized Caching mechanism).KCache利用控制器获取各路由器上内容的请求次数以及网络状态信息,并据此计算得到若干个缓存节点,向其发出主动缓存内容的指令.仿真实验表明:与传统内容中心网络缓存机制相比,KCache可有效提高缓存内容命中率,降低获取内容的跳数以及降低平均请求时延.

Internet usage has dramatically transformed from host-centric end-to-end communication to content retrieval. In order to adapt to this change,content centric networks have been proposed. In-network caching has been viewed as an attractive feature of CCN because it can reduce network traffic and user access latencies. In order to reduce the cache redundancy as well as increase the cache hit ratios in content-centric networks,we proposed the K-means algorithm utilized Caching Mechanism（ KCache） mechanism. KCache utilizes controllers to obtain the number of content request of each router and collect network status. Then the controller will calculate several caching nodes according to the upload information and issues a proactive caching content instruction to these nodes. Simulation results show that KCache can effectively increase the cache hit ratios,reduce the content acquired hops and reduce the average request delay compared with the traditional routing algorithm in CCN.