Research On Named Look-up Method For Named Data Networking

Named Data networking is one of the concrete implementation plan of the InformationCentric Networking.The NDN evolved from traditional data transmission in IP network to the information interaction mode based on utilize the name of the content itself.Thus it overcomes the problems of network address translation and lack of security.Therefore,NDN is regarded as a potential network architecture of the future Internet.NDN architecture has the advantage of using the data content name instead of the previous host addressing.It can reduces the burden of data bandwidth consumption to make fast response time.The NDN architecture is no longer focused on the location of the information,it focus on information content.With the recent explosive growth in Internet communications,it is very important to find research on data names in NDN.In this paper,we study the NDN data name look-up in two optimization problems: the longest prefix matching search algorithm in FIB(Forwarding Information Base),the precise optimization search algorithm in the CS(the Content Store)and the PIT(Pending Interest Table).And the concrete research content is as follows:This paper introduces the design principle of naming data network architecture,packet routing and routing and naming mechanism.The two optimization problems of packet forwarding are discussed in detail.Multilevel count bloom filter-based the longest prefix matching search algorithm in FIB is proposed.The scheme bases on the counting bloom filter,which designs the diminishing length of multilevel bloom filter to search.And it adds the concept of the fingerprint.The traditional counting cell is divided into four parts in bloom filter,which is used to store the record of each string of fingerprint information.The scheme compared with DLCBF scheme,the conflict probability decreases by 15% on average,and by setting the label function,effectively guarantee the higher search speed.Sum-up counting Bloom filter-based the exact matching algorithm in CS and PIT is proposed.The scheme adds the summary table and confirm table to effectively avoid the past bloom filter characters divided fault situation.Experimental results show that the efficiency of the proposed method compared with conventional methods,which the storage time shortened by 10%,the conflict rate is better than the traditional scheme.