Efficient Regular Expression Signature Generation for Network Traffic Classification

Regular expression signatures are most widely used in network traffic classification for trusted network management. These signatures are generated by the sequence alignment of the traffic payload. The most commonly used sequence alignment algorithm is Longest Common Subsequence (LCS) algorithm which computes the global similarity between two strings but it fails in consecutive character matches. This paper presents a new divide and conquer alignment algorithm for generating regular expression signature by rewarding contiguous character matches. The results indicate that the sequence alignment algorithm that used is the space efficient way and the algorithm outperforms LCS in terms of efficiency and accuracy.