摘要
在软件漏洞挖掘领域,Fuzzing测试是使用最广泛、最有效的方法之一。传统Fuzzing测试方法存在工作效率低、盲目性强等不足。该文提出一种样本集精简算法和一种加权的测试时间模型,能够在保证代码覆盖率不变的情况下减少测试样本的数量,同时使优质的样本得到更多的测试时间片;设计了一种基于污点传播的异常分析方法,可评估异常信息的危害程度,有助于提高漏洞分析的效率。实验结果表明:与Peach实验进行对比,该文提出的方法有效地改进了传统的Fuzzing测试方法。
Fuzzing testing is one of the most widely used and most effective methods for vulnerability detection. However, the traditional fuzzy analysis method is inefficient and works blindly. This paper describes a refining method that reduces the test sample size with the same code coverage. A weighted testing time model is used to give the better sample more time. A taint based exception analysis method is used to evaluate the severity of exceptions and to improve the vulnerability analysis efficiency. Comparisons with Peach show that this method improves the traditional fuzzy analysis method.
出处
《清华大学学报（自然科学版）》
EI
CAS
CSCD
北大核心
2016年第5期478-483,共6页
Journal of Tsinghua University(Science and Technology)
基金
国家自然科学基金资助项目(61502536)
河北省高等学校青年拔尖人才计划项目(BJ201414)
关键词
模糊测试
精简集
漏洞分析
Fuzzing
refining set
vulnerability analysis
