New types of malware are emerging at concerning rates. However, analyzing malware via reverse engineering is still a time-consuming and mostly manual task. For this reason, it is necessary to develop techniques that automate parts of the reverse engineering process and that can evade the built-in countermeasures of modern malware. The main contribution of this paper is a novel method to automatically find so-called Points-of-Interest (POIs) in executed programs. POIs are instructions that interact with data that is known to an analyst. They can be used as beacons in the analysis of malware and can help to guide the analyst to the interesting parts of the malware. Furthermore, we propose a metric for POIs , the so-called confidence score that estimates how exclusively a POI will process data relevant to the malware. With the goal of automatically extract peers in P2P botnet malware, we demonstrate and evaluate our approach by applying it on four botnets (ZeroAccess, Sality, Nugache, and Kelihos). We looked into the identified POIs for known IPs and ports and, by using this information, leverage it to successfully monitor the botnets. Furthermore, using our scoring system, we show that we can extract peers for each botnet with high accuracy.
翻译:然而,通过反向工程分析恶意软件仍是一项耗时且主要是人工的工作。 因此,有必要开发一些技术,使反向工程过程的某些部分自动化,并且可以回避现代恶意软件的内在反制。 本文的主要贡献是,一种在已执行的程序中自动找到所谓的“利益点”的新颖方法。 污染物是与分析员所知数据互动的指示。 它们可以用作分析恶意软件的灯塔,并能够帮助分析员了解恶意软件的有趣部分。 此外,我们建议了对 POI 的衡量标准,即所谓的信任评分,即估计一个POI将专门处理与恶意软件相关的数据。为了在P2P Bottnet软件中自动提取同侪,我们通过将它应用到四个机器人网络(ZeroAcess、Sality、Nugache和Kelihos)来展示和评估我们的方法。 我们用已知的 POIS 来查看已知的IP和端口端点。 此外,我们用这个信息,即所谓的信任评分系统来成功监测我们的磁网。