Research Of Cooperative Caching Strategy In Named Data Networking

With the development of communication technologies and electronic hardware,number of communication terminals and contents in the Internet increases in a geometrical ratio.The researchers proposed a content-centric network architecture called NDN(Named Data Networking).The advantage of NDN is that each router node can store returned data for clients' subsequent requests.People have made a lot of researches in caching strategy of NDN.Because of decentralization of router nodes location and the limitedness of cache space,these caching strategies cause the problem of low cache space utilization.This thesis improves the caching strategy and cache replacement strategy,to increase cache space utilization and the hit rate of clients' requests in the network.Firstly,this thesis introduces background and significance of the research.In addition,introducing previous caching strategies and caching replacement strategies of NDN to analyze advantages and disadvantages of these strategies.Secondly,the current cache decision strategies do not treat the contents differently in the cache,which makes the data in the network high redundancy.In this thesis,collaborative cache strategy(Co-Po)based on popularity prediction is proposed.The copies of data with high popularity are stored in multiple nodes on the returned path.Compared with the connectivity of nodes on returned path and that of its neighborhood nodes,the copies of data with low popularity are stored in the well-connected nodes.So that when clients request contents,a higher hit rate can get in NDN.In addition,when the router node storage space is full,it needs to replace the content.The existing replacement strategies always consider the single factor.So these strategies can't evaluate the value of dynamic contents in nodes accurately.Aiming at the problem,this thesis proposes a cache replacement strategy(Fr-Re)based on frequency and recency.Which considers recency and frequency of the content in the node cache,in the meantime,the routing hop count is used as the cost index in the network.Comprehensive the factors discussed above,this thesis evaluates the value of the contents in the router nodes,and sorts the contents according to the value.When the node needs to replace content,it replaces the lowest-value content to improve utilization of the cache space.Finally,this thesis chooses the ndnSIM simulation module based on NS-3 platform to simulate and give the analysis for the Co-Po cache decision strategy and Fr-Re replacement strategy.The simulation results show that the Co-Po and Fr-Re strategies have better network performance compared with the existing strategies in terms of hit rate,hop count and delay.