一种基于两级缓存的协同缓存机制

信息中心网络（information-centric networking，简称ICN）将网络通信模式从当前的以地址为中心转变为以信息为中心.泛在化缓存是ICN重要特性之一，它通过赋予网络任意节点缓存的能力来缓和服务器的压力，降低用户访问延迟.然而，由于缺少内容热度的分布感知，现有ICN缓存策略仍存在缓存利用率较低、缓存位置缺乏合理规划等问题.为了解决这些问题，提出一种基于两级缓存的协同缓存机制（a cache coordination scheme based on two-level cache，简称CSTC）.将每个节点的缓存空间分为热度感知和协作分配两部分，为不同热度的内容提供不同的缓存策略.同时，结合提出的热度筛选机制和路由策略，降低了缓存冗余，实现了缓存位置优化.最后，基于真实网络拓扑的仿真实验表明，CSTC在次热门内容缓存数量上提升了2倍，缓存命中率提升了将近50%，且平均往返跳数在多数情况下优于现有On-path缓存方式.

Cache Coordination Scheme Based on Two-level Cache

Information-centric networking (ICN) transforms the network communication mode from the current host-oriented mode to an information-oriented one. Ubiquitous in-network caching is one of the significant features of ICN, which can effectively alleviate server pressure, as well as decrease the user access latency by allowing any nodes in network to cache. However, due to the lack of distribution awareness of content popularity, there are still many problems with the state-of-the-art ICN caching schemes, such as low cache utilization and lack of reasonable planning of cache location. This study proposes a cache coordination scheme based on two-level cache (CSTC) to solve these problems. The content store (CS) of each node is divided into two parts:popularity perception and collaboration allocation. Different caching strategies are applied to cached content with different popularity. At the same time, combined with the popularity filtering and routing mechanism, this scheme reduces cache redundancy and optimizes cache location. Finally, simulation experiments based on real network topology show that CSTC has doubled the number of cached secondary popular content. The cache hit ratio has increased by nearly 50%, and the average round-trip hop count is superior to the existing on-path caching method in most cases.