Comparing the effectiveness of commercial obfuscators against MATE attacks
Ramya Manikyam, J. Todd McDonald, William Mahoney, Todd R. Andel, Samuel H. Russ
University of South Alabama University of Nebraska at Omaha
阅读操作
确认中在文库中上传 PDF 后可生成中文音频讲解。
摘要与影响
The ability to protect software from malicious reverse engineering remains a challenge faced by commercial software companies who invest a large amount of resources in the development of their software product. In order to protect their investment from potential attacks such as illegal copying, tampering, and malicious reverse engineering, most companies utilize some type of protection software, also known as obfuscators, to create variants of their products that are more resilient to adversarial analysis. In this paper, we report on the effectiveness of different commercial obfuscators against traditional man-at-the-end (MATE) attacks where an adversary can utilize tools such as debuggers, disassemblers, and de-compilers as a legitimate end-user of a binary executable. Our case study includes four benchmark programs that have associated adversarial goals categorized as either comprehension or change tasks. We use traditional static and dynamic analysis techniques to identify the adversarial workload and outcomes before and after each program is transformed by a set of three commercial obfuscators. Our results confirm what is typically assumed: an adversary with a reasonable background in the computing disciplines can both comprehend and make changes to any of our completely unprotected programs using standard tools. Additionally, given the same skill set and attack approach, protected programs can still be probed to leak certain information, but none could be successfully altered and saved to create a cracked version. As a contribution, our methodology is unique compared to prior studies on obfuscation effectiveness in that we categorize adversarial skill and delineate program goals into comprehension and change ability, while considering the load time and overhead of obfuscated variants.
逐年被引趋势
关键指标
同类平均 = 1
同领域 · 同年份 · 同类型
Google Scholar 与 OpenAlex 的被引统计范围不同,数值存在差异属正常。
AI 辅助阅读
依据:摘要
可就本文提问;依据不足时会说明。
学术脉络
学科主题
计算机 / AISecurity and Verification in Computing
Physical Unclonable Functions (PUFs) and Hardware Security · Digital Rights Management and Security