CRBP-OpType: A constrained approximate search algorithm for detecting similar attack patterns

Abstract

Misuse-based Intrusion Detection Systems (IDS) have difficulties in detecting new attacks originating from the old ones, since they usually use exact search algorithms to detect attack patterns. To solve this problem, we can use approximate search. However, approximate search algorithms generate too many false positives. In this paper, we introduce constraints on the types of edit operations in the approximate search. Our algorithm applies an extension of the concept of Row-Wise Bit-Parallelism (RBP), which makes it highly efficient. The time complexity of the algorithm is at worst the same as that of the unconstrained approximate search algorithm based on the same RBP concept; otherwise, in most cases it is better than the unconstrained RBP. Experimental results show that our constrained approximate search algorithm produces a smaller number of false positives compared to the unconstrained RBP search algorithm, without reducing accuracy of search.