Research On The Method Of Content Name Lookup For Information-Centric Networking

Information-centric networking is a future Internet structure which is proposed to meet the growing information needs of users and respond the challenges of the current Internet.It abandons the traditional IP-based network architecture,adopts content as the core of network structure and services,and takes the content name as the identity of network transmission.However,due to the variety of content name naming methods,the structure of the name is complex,and the length of the name is not limited,it has become a challenging task to achieve accurate and efficient lookup of the content name in a large-scale routing table.In the name lookup process,the lookup is performed in three tables on the routing node: Forwarding Information Base,Content Store and Pending Interest table.This paper mainly studies the optimization problem of content name lookup on the three tables of Information-centric network,the specific content is as follows:1.A name lookup structure based on learned Bloom filters is proposed.This method is designed for the optimization problem of name lookup based on the longest prefix matching algorithm in the Forwarding Information Base table.We use recurrent neural networks(RNNs)to build a learned model to pre-find names,and then use a backup Bloom filter to eliminate false negatives generated by the learned model to improve the accuracy of the name lookup.Compared with the traditional Bloom filter,the proposed structure has a lower false positive rate and a higher lookup speed.2.A three-level name lookup structure based on deep Bloom filter is proposed.This structure optimizes the exact lookup of content names in the Content Store and Pending Interest table.The first level sets an original filter to pre-filter the names.In the second level,a Long-Short Memory neural network(LSTM)with Gate Recurring Unit(GRU)is introduced to construct the deep learned model,send the first-level filtered names to the learned model for precise lookup to determine the name and find the port corresponding to the name.The third-level backup filter is used to eliminate the false negatives.Experiments show that the optimized precise lookup method can significantly increase the lookup accuracy,reduce memory consumption,and improve the efficiency of precise lookup.This paper focuses on the lookup of content names in ICN,and optimizes the problems in the two lookup methods: longest prefix matching and precise lookup,this paper proposes a learned Bloom filter structure and a three-level lookup structure based on deep Bloom filter for name lookup.The experimental results show that these name lookup methods proposed in this paper are superior in terms of speed,accuracy and memory consumption.