Research On In-Network Cache Optimization For Content Centric Network

With the constant development of network technology and the popularization of informationization,as the requirement of getting content type,accessing pattern and data size continues to change,the network architecture are also continue to change,including both framework revolution and architecture innovation.The content centric network(CCN),which overturned the traditional Host-To-Host network communication mode,has attracted widely attention of researchers.The content-centric communication mode in CCN is achieved by distributed in-network cache and requester-driven routing mechanism.The in-network cache feature as an integral part of CCN,caching strategy directly affects the network performance.While the default everywhere-caching strategy of CCN couldn't effectively utilize in-network cache resources,therefore do not make full use of cache resources to achieve the purpose of efficient data transmission.This paper's purpose is to improve the content transmission efficiency of CCN by optimizing network cache model and cache placement/replacement policy.The research involves single node replacement algorithm,cooperative cache,routing cache combination and mobility support.Firstly,the everywhere-caching attribute of CCN results in cache redundancy and the independent node cache reduces the hit-ratio of contents with different popularity.To deal with this problem,on the basis of the “Edge theory”,which means caching high popularity contents in edge nodes closer to requester,this paper has presented an optimal cache placement policy based on content popularity.This policy could build content popularity statistic model from the records of users' request,which is used to achieve the optimal placement and to improve utilization of cache.An optimal cache placement model also had been put forward to verify the presented policy.Secondly,aimed to solve the problems of single node cache policy's limitation and lacking the ability to share other nodes' resources,this paper has presented a low-complexity nodes cooperation caching policy,which consists of two parts: cache placement and cache replacement.On the arrival of new contents,if the content store is full then the contents(cached)already cached in neighbors will be replaced,otherwise the contents(arrived)notcached in neighbors will be cached.By these means transmission efficiency could be improved,because more requests could be fulfilled in the distance of one hop.Also,contents diversity is promoted.Meanwhile,in order to further improve the space utility rate of limited node cache and the efficiency of transmission,this paper has presented a cache optimization solution by combining routing and caching,namely the weighted Multi-Hash Routing caching mechanism based on content popularity.This mechanism could dynamically adjusts the number of cache copies in the network,and deploy these copies among nodes of routing paths by improved Multi-Hash function and edge caching theory,therefore allows nodes to locate requested contents more accurate and faster.Also,taking the content popularity into account could improve the transmission efficiency.Finally,to deal with the ever-changing topology and cache contents of mobile CCN and to ensure longstanding data reliability,this paper has presented an algorithm of Triplication Cache Partitioning.Based on the nodes' location,this algorithm could set the threshold limit of communication distance between nodes,therefore using an optimal scheme to divide the whole network into triplication-nodes group without public node,and accomplish real-time cache content synchronization in each group.Experiment shown the group algorithm achieved good performance under the condition of high node density and low cache space than default cache strategy in CCN.