Dynamic Attack Tree and Its Applications on Trojan Horse Detection

Abstract

Attack tree generating algorithm has been studied thoroughly. However, the method for adjusting the attack tree with ever-increasing needing has not been presented. A novel Trojan horse detection approach is presented in this paper. Original attack tree is modeled on the basis of APIs that Trojan horse often used. Matching the set of APIs got by statically analyzing the PE file with the original attack tree, to predict malicious actions may appear in the implementation of the PE file and estimate whether the PE file is a Trojan. The attack tree can adjust with word analysis after analyze the detection results to improve the detection ability. The experiment shows that this approach can detect known and unknown (upgraded) Trojan horse effectively. Moreover, this is the first time present the attack tree's dynamically adjustment and this has for-reaching significance for the attack tree and Trojan horse detection.