A generalized adversary decision algorithm and analytic solution methods for advise models

Quantitative security metrics are becoming increasingly important to system administrators. ADVISE generates quantitative security metrics by combining a system vulnerability graph with an adversary profile through an adversary decision algorithm. Previously, the decision algorithm placed restrictive assumptions on the adversary profile, and simulation was the sole solution method for ADVISE models. In this thesis, the decision algorithm is generalized while simultaneously improving its performance by incorporating theory from discrete-time Markov games. Furthermore, by exploring the state-space and generating the transition probability matrix, numerical solution methods may be applied to solve ADVISE models. Identifying key properties allows the models to be tested for compatibility with alternative solution methods from the literature, enabling additional metrics for ADVISE models. Finally, the performance of simulation is improved significantly by introducing decision caching. Together these accomplishments expand the number of quantitative security metrics and solution methods available to ADVISE models while lifting restrictions on the adversary profile and improving performance.

Name thesis:degree_name

M.S.

Level thesis:degree_level

Thesis

Discipline thesis:degree_discipline

Electrical & Computer Engr

Grantor

University of Illinois at Urbana-Champaign

Year dc:date

2012

Author dc:creator

• Ford, Michael

Contributors dc:contributor

• Sanders, William H.

• Security models
• quantitative metrics
• state space exploration
• numerical techniques

Statement dc:rights

• Copyright 2012 MICHAEL D. FORD

Language dc:language

en

Handle dc:identifier

http://hdl.handle.net/2142/31008

OAI identifier oai:identifier

oai:www.ideals.illinois.edu:2142/31008