Scalable and Efficient Pathsensitive Analysis Technique Scanning Many Types of Vulnerability


Dongok Kang and Minsik Jin, R&D Center, Repulic of Korea


The goal of this paper is to present an efficient and effective path-sensitive analysis technique for many types of security vulnerability. We propose two analysis techniques. The first is a scalable path-sensitive analysis technique for security vulnerability with high precision and recall. Our strategies are to allow flexible design of path state and to make an effective path navigation heuristic which achieves both scalability and high recall. Experimental results show that a vulnerability scanner implemented through this technique get precision 100% and recall 93% on OWASP Benchmark. The vulnerability scanner is able to analyze 1 million lines of code. The second is a pre-analysis technique to improve the efficiency of the above analysis technique. The pre-analysis technique improves the path navigation by using an additional cheap anlysis. Despite the additional cost, experimental results show that the total analysis time is reduced by 2.5 times. Simultaneously recall of the analysis is improved by the pre-analysis technique.


Secure coding, Security, Static analysis, Vulnerability scanner, Summary-based, Path-sensitive, Information flow Analysis, Pre-analysis

Full Text  Volume 7, Number 8