Applying Standard Network Centrality Measures to Analyze Error Propagation and Measure the Security of a Software System

Software systems tend to be large and complex, consisting of multiple components. Current security assessment methods are predominantly qualitative and do not reveal information pertaining to how flaws propagate from one component to another. Furthermore, they do not provide information regarding the interdependencies of the flaws on one another and between components. In order to measure the overall security of a software system that consists of multiple components, the researcher used a weighted directed graph to model the system to depict the interrelations and effects of software security flaws existing within its components. Standard centrality measures of the graph, including closeness, degree, betweenness, and eigenvector, are examined to assess the importance and activity of a component in the propagation of software flaws in multiple components across a software solution. Calculating the centralities of the nodes in the graph of the software system provided information to determine how influential each node is within the system and how significant a node is in propagating security flaws.

Author

• Smyre, Christina

Language

• en

Keyword

• centrality
• software
• flaws
• network
• propagation
• dependency

Date created

• 2011

Type of Work

• Dissertation

License

• All rights reserved

Rights statement

• In Copyright

GW Unit

• Computer Science

Degree

• Ph.D.

Advisor

• Martin, Dianne

Committee Member(s)

• Vora, Poorvi
• Williams, Yul
• Shaw, Gregory
• Heller, Rachelle

Persistent URL

•  https://scholarspace.library.gwu.edu/etd/4q77fr45d