基于缓冲区溢出的数据驱动型漏洞检测系统总体设计与实现

Design and Realization of Data-driven Vulnerability Detection System Based on Buffer Overflow

针对缓冲区溢出漏洞提出了一种检测系统并加以实现.检测系统的原理是在原有的语法分析和词法分析的基础上,采用区间运算和改进的Wagner算法等方法,实现了语义分析更深层次上的检测.同已有的检测方法相比,该检测方法实现了对缓冲区溢出漏洞更准确的检测,具有较低的误报率和很好的运行性能.

For buffer overflow vulnerabilities, a face detection method is put forward and realized. This detection method is based on the original syntax and lexical analysis, tusing interval arithmetic and improved algorithm of Wagner and other methods, a deeper semantic analysis of the detection is achieved. Compared with existing test methods, the more accurate detection method enables buffer overflow vulnerability detection, has a lower false positive rate and very good running properties.