提出了一种基于流行度的概率存入校验机制(PCS-CP),根据接收内容的流行度,概率抽取内容校验,并只存入校验通过内容,确保节点的有限计算资源尽可能服务于用户关注内容,无须校验网内命中内容,降低了校验开销.由于PCS-CP机制只有与特定缓存策略配合才能最大化效用,进而提出了一种基于缓存更新时间的网内缓存策略,对网内副本冗余进行优化控制,有效提升了PCS-CP机制的校验效果.数值结果表明,与命中校验机制相比,PCS-CP可有效降低网内校验次数,有效防御内容污染攻击.
This article proposed a checking mechanism, named as probabilistic checking before storing based on content popularity (PCS-CP). The design of PCS-CP includes two aspects, one is the node which should check the received content randomly according to its popularity, and the other is the node only stores the legitimate content. On the one hand, PCS-CP can ensure the limited computing resource of node to serve the high popularity content as much as possible. On the other hand, PCS-CP guarantees the authenticity of cached content, and then reduces the computation overhead of node. Because only co- operating with certain caching policy, PCS-CP can maximize its effectiveness. The in-network caching strategy based on cache update time (ICS-CUT) was further proposed to optimize the copy redundancy in network. It is shown that, comparing with checking on hit mechanism, PCS-CP can effectively reduce the average amount of checking in network and well defense the content pollution attack.