Abstract
String matching is the most critical operation in network intrusion detection Systems (NIDS) [1]. This paper proposes a novel memory-efficient string matching algorithm that only requires around 2% of the memory utilized in Aho-Corasick algorithm [2] but has more than 4 times the throughput of stateof- the-art algorithm with very limited memory resource. The proposed algorithm is flexible to fit different resource constraints and performance requirements. © 2008 IEEE.