摘要
在软件定义网络(SDN)和内容中心网络(CCN)融合架构下,为了充分利用控制层对网络拓扑和缓存资源的全局感知,在全网中实现缓存资源的优化使用,提出了一种集中控制的缓存决策优化方案。在该方案中,应用粒子群优化算法(PSO)并且根据节点边缘度、节点重要度以及内容流行度对缓存资源和内容进行集中缓存决策,使得内容在不同的节点进行合理的缓存。仿真结果表明,通过评估缓存大小对缓存性能的影响,PSO缓存决策方法取得了比LCE、PROB缓存决策策略更优的缓存命中率和路径延展率,明显降低了缓存节点的缓存替换数,使得缓存达到了整体缓存优化。
In the soRware defined networking and content centric networking integration framework, a cache optimi- zation scheme of centralized control was proposed to make full use of global awareness of the control layer for net- work topology and cache resources, and achieve the optimal use of caching resources in the entire network. In this scheme, the particle swarm optimization （PSO） was applied for centralized cache decision-making depending on node edge degree, node importance degree and content popularity. Therefore, the contents were cached in different nodes reasonably. Simulation result shows that the proposed PSO caching scheme achieves larger cache hit rate and lower path stretch than that of LCE scheme, PROB scheme by evaluating the impact of cache size and content popularity on the caching performance. It also shows that PSO scheme significantly reduces the number of cache nodes to replace the cache, making the cache to achieve the overall cache optimization.
出处
《电信科学》
北大核心
2017年第5期12-20,共9页
Telecommunications Science
基金
国家自然科学基金资助项目(No.61371087
No.61531013)
教育部-中国移动科研基金资助项目(No.MCM20150102)~~
关键词
软件定义网络
内容中心网络
缓存
粒子群优化
节点边缘度
节点重要度
内容流行度
software defined networking, content centric networking, caching, particle swarm optimization, nodeedge degree, node importance degree, content popularity
