An Improved Attack Path Discovery Algorithm Through Compact Graph Planning

Abstract

Attack path discovery plays an important role in protecting network infrastructure. Unfortunately, previous attack path discovery algorithms are difficult for applying in reality because of the high computational complexity problem. To achieve effective attack path discovery, we proposed a compact graph planning algorithm to incorporate goal states related information into attack path discovery. Our model extracts goal states related information by calculating closure of goal states, and then construct planning graph structure given the closure, after which the backward search algorithm is used to extract the attack path solution. The experiments were done on the typical enterprise network, comparing the effectiveness of attack path discovery algorithms with existing known methods. The result turns out that our proposed compact graph planning algorithm shows great improvement in discovering attack paths.