基于跨层信息的信息中心网络缓存算法研究

摘要

To improve the performance of information-centric networks ( ICN) , the optimal design of ICN caching algo-rithms was studied.In view of the fact that current ICN caching algorithms have the drawbacks of high content re-dundancy and low cache hit rate, a novel ICN caching algorithm based on cross-layer design, called CRCache, was proposed.The CRCache algorithm caches each specific content on its matchable router according to the correlation of the network layer’ s information and the application layer’ s information.The cross-layer design can eliminate un-necessary redundant contents in ICN, improve the cache hit rat and reduce the network traffic.A large scale and real traces-driven simulation with an underlying real Internet topology in China was designed and conducted, and the results show that the caching performance of the CRCache algorithm exceeds current schemes in the respects of hit rate, traffic reduction and server load.%为了提高信息中心网络（ ICN ）的性能，研究了ICN缓存算法的优化设计。针对现有的缓存算法冗余度高、命中率较低的问题，提出了一种基于跨层设计的全新的ICN缓存算法CRCache，该算法根据网络层信息与应用层信息之间的相关性，有选择地将内容缓存在与其相匹配的路由器上。通过这种跨层的缓存设计，CRCache缓存算法能够消除网络中不必要的冗余内容、提高缓存命中率、减少网络流量。设计并实现了基于真实用户访问日志与真实中国网络拓扑的大规模仿真实验，通过实验证明，该算法在缓存命中率、网络流量减少率、服务器负载等指标上均优于现有的缓存算法。