Efficient Name Lookup Scheme Based on Hash and Character Trie in Named Data Networking

Abstract

Name lookup, as a key function of routers for forwarding and filtering packets in NDN, has confronted serious memory efficiency problem. To solve the problem, this paper presents an efficient NDN name lookup scheme -- Split Name character Trie(SNT). It first decomposed a name into components, which were then hashed to construct a trie. After that, the trie would be split into two smaller ones. When operating name lookup, a name would be split into two parts to conduct lookup on the two tries independently. Experimental results show that SNT supports longest prefix match very well. It can reduce 30% ~ 38% memory cost by component hash and can reduce 30% ~ 33% memory cost further by splitting trie. Despite of the extra memory cost of maintaining hash table, it still improves general memory efficiency by 23% ~ 49%.