内容分发网络中基于相关内容吸引的缓存算法

Cache Algorithm Based on Related Content Attracting in Content Delivery Network

内容分发网络中基于内容名的缓存算法会导致路由表规模随网络增长而膨胀,将严重影响网络路由效率和性能。针对该问题,提出一种基于相关内容吸引的节点缓存算法。利用本地缓存算法,通过节点已缓存内容对其他内容的吸引作用吸引主要特征内容,排斥具有次要特征内容,将缓存中不同特征内容的数量差异进行放大,使缓存内容表现出明显稳定的内容特征。同时设计相关内容生存时间相互增强的缓存策略,以减少路由通告信息量,提高内容分发网络的路由能力。实验结果表明,该算法在有效解决路由问题的同时,能增强缓存内容稳定性,提高路由可信度。

The existing content cache algorithm of Content Delivery Network（CDN）leads to the expansion of routing table with the network increasing,which will impair the routing efficiency and network performance. Therefore,based on related contents attracting,a related contents attracting algorithm is proposed. With the effect of attracting similar contents cached in other near nodes,for the purpose of apparently stable featured contents of nodes cached,the algorithm attracts major characteristic contents,rejects secondary feature contents,and enlarges the difference of different characteristic content. It also gathers the related contents on the same nodes via the mutual attraction with same contents feature,which facilitates the cache contents feature abstraction. Meanwhile,the strategy of lifetime increasement between contents with main feature is designed to deduce the routing advertisement and improve the routing scalability. Experimental results show that the proposed algorithm can reduce the update frequency of cache content,and improve the routing reliability.