Using Attack Surface Entry Points and Reachability Analysis to Assess the Risk of Software Vulnerability Exploitability

An unpatched vulnerability can lead to security breaches. When a new vulnerability is discovered, it needs to be assessed so that it can be prioritized. A major challenge in software security is the assessment of the potential risk due to vulnerability exploitability. CVSS metrics have become a de facto standard that is commonly used to assess the severity of a vulnerability. The CVSS Base Score measures severity based on exploitability and impact measures. CVSS exploitability is measured based on three metrics: Access Vector, Authentication, and Access Complexity. However, CVSS exploitability measures assign subjective numbers based on the views of experts. Two of its factors, Access Vector and Authentication, are the same for almost all vulnerabilities. CVSS does not specify how the third factor, Access Complexity, is measured, and hence we do not know if it considers software properties as a factor. In this paper, we propose an approach that assesses the risk of vulnerability exploitability based on two software properties - attack surface entry points and reach ability analysis. A vulnerability is reachable if it is located in one of the entry points or is located in a function that is called either directly or indirectly by the entry points. The likelihood of an entry point being used in an attack can be assessed by using damage potential-effort ratio in the attack surface metric and the presence of system calls deemed dangerous. To illustrate the proposed method, five reported vulnerabilities of Apache HTTP server 1.3.0 have been examined at the source code level. The results show that the proposed approach, which uses more detailed information, can yield a risk assessment that can be different from the CVSS Base Score.