Research On The Detection And Defense Strategy Of Cache Pollution In Named Data Network

Because of the rapid expansion of Internet users,Internet traffic has also doubled.The host-centric network communication model adopted by current TCP/IP architecture no longer satisfies the fast and near-term acquisition of massive information,resulting in the reduction of network resource utilization and transmission efficiency.In order to improve the efficiency of content delivery,breaking the end-to-end architectures limitiaon,serveral content distribution technologies are merging,such as P2P and CDN which work as overlay above TCP/IP.A network level architecture which provides primary support for saclable and efficient content retrieval should be proposed.One of these kinds of network level architecture is Information-Centric Network(ICN).Named Data Network(NDN)is one of the most promising ICN architectures.The ubiquitous cache in Named Data Network brings new security threats,especially cache pollution attack.Different from flooding-based denial-of-service attack,cache pollution attack manipulates the content in nodes' cache store by controlling the distribution of incoming Interest.According to the requests,cache pollution attack can be divided into False-Locality Attack(FLA)and Locality-Disruption Attack(LDA).This kind of attack is concealment and latent,which can do great harm to NDN.In this paper,the harmfulness and characteristics of cache pollution attack are study through the modeling of cache pollution attack.A detection and mitigation method are proposed based on cached content identification.The defense strategy is divided into two stages:detection stage and defense stage.In detection stage,Decision-tree-based Support Vector Machine is used to classify cached content.In defense stage.Two corresponding defense strategies for two kinds of cache pollution are proposed.Through simulation,the defense strategies perform well in the mitigation of cache pollution attack.