Thesis (D.Eng.)--Chulalongkorn University, 2010

Because of the complexity of software, vulnerability are unavoidable. Vulnerability is a common path for attacking the system. Straightforward protections are applying system patches. Due to the complexity of the software and limited resource in administrative work, system administrator has difficulties in applying remediation. To maximize work performance and security level of the system with limited administrative resource, vulnerability needs to be prioritized. This work proposed the analysis of vulnerability relevancy using metrics based on global public information. Vulnerability relevancy is defined based on vulnerability lifecycle and public information acquisition factors. Vulnerability Relevancy Quantification Framework is proposed using web data mining concept with ontology. The experiments are conducted on top of severe vulnerability announcements from established security organizations. The results show that top severe vulnerabilities from the published list have different level in relevancy depending on vulnerability age and availability of remediation process. The highly ranked vulnerability is related to clientside software and common software libraries.