An iterative pattern mapping for parallel string matching architecture in intrusion detection systems

Abstract

This paper proposes an algorithm that maps target patterns onto parallel string matching architectures in intrusion detection systems (IDS). In the proposed iterative pattern mapping, the sets of patterns that are mapped onto string matchers are sorted in ascending order of the average pattern length in each turn. By mapping a set of patterns for a string matcher onto the other string matchers repeatedly, the required number of string matchers is reduced. Therefore, the proposed iterative pattern mapping minimizes the total memory requirement for parallel string matching architecture.