Study On Caching Policies And Security Technologies For Content Centric Networks

As a hot topic of next generation Internet architecture,content centric networking(CCN)has attracted many researchers' attention.Using distributed caching mechanism and name routing,CCN can realize content-oriented data transmission regardless of the location of stored data.It efficiently solves many thorny problems caused by the traditional host-based communication mode,such as bandwidth competition,traffic congestion and redundant transmission.As the key to CCN,caching mechanism can give rise to advantages in many aspects,but in the meantime will lead to some problems.On one hand,ubiquitous caching enables users to obtain contents at the network edge so as to reduce response time,relieve network congestion and improve utilization of network resources.On the other hand,as a public platform for data exchange,caching mechanism brings about some security threats,including caching pollution and caching privacy leakage.For this reason,the research on caching mechanism and its security problems in CCN has great theoretical and practical significance.The research resluts will be of benefit to the performance improvement of CCN and its pace towards coming into practice.Taking into account of above considerations,the dissertation focuses on caching policies and caching security technologies of CCN.The research topics include cache replacement policy,cache decision policy,cache pollution attack protection and cache privacy protection.The innovations of the research are as follows.(1)Aiming at improving in-network cache utilization,a cache probability replacement policy,named PP policy,is proposed based on content popularity.It manages to reduce the network storage redundency by increasing the caching proportion of non-popular contents.Based on node caching modelling and theoretical analysis of policy performance,numerical resluts are made for request miss probability and avarage hit disitance of both single node and the whole network.The effect of reducing network storage redundency is also explored.(2)A cache decision policy based on probabilistic caching with evicted copy up,named PCECU policy,is proposed.It can increase the network caching time of contents using the mechanism of moving copy to upstream node,and increase caching proportion of popular contents at the network edge using probabilistic caching mechanism.It will significantly improve the network performance by taking into account two design goals,i.e.to improve hit probability of content request and to decrease the retrieving delay of popular contents.An accompanying innovation work in this respect is to propose a generalized theoretical analysis method for deriving layered hit probability by introducing the concept of caching characteristic time.It gets rid of the constraint of popularity parameter ?(29)1,which is the limitation of traditional analysis methods.Using the proposed method,the hit probability and its influencing factors of PCECU are evaluated for UGC and VOD services,and the network performance improvement is verified owing to the mechanism of probabilistic caching with evicted copy up.(3)To defend against cache pollution attacks,a traffic limiting mechansim is proposed based on trusted judgment of face abnormality.It detects pollution attack category by face abnormality and limits traffic on the abnormal face according to its request hit probability.The mechanism can defend against both locality-disruption and false-locality attacks.In addtion,through judging the occurrence probability of false-locality attack,the abnormal false-locality attack can be distinguished from the normal flash crowd by trusted judgment of abnormality status.The mistaken restraint can be avoided on the legitimate behavior of flash crowd.Compared with existing approaches,the proposed mechanism exhibits prominent superiority.(4)To solve the cache privacy leakage problem,investigation is made from two different perspectives.One is caching policy design,from which a cache protection policy is proposed based on recent visiting information and evicting copy up mechanism.The other is content request filtering,from which a cache privacy protection mechanism is proposed based on dynamic address mapping.The former approach distinguishes different user requests with recent visiting information and blurs content caching time in the network with evicting copy up mechanism.Through setting resonable evicting probability,it can maintain low privacy leakage probability while achieving high request hit probability,leading to a nice balance between cache utilization and privacy leakage in CCN.The latter approach distinguishes user's first request by Bloom filter based dynamic address mapping.It can provide strong cache privacy protection by adding additional computational and storage cost and has good deployment feasibility through storage optimization.Embodying current mainstream ideas of cache privacy protection,these two proposed approaches have important reference value to secured CCN design.