Modeling and analysis of trustworthy systems using extensions of network reliability

Nguyen, Hoang Hai

Permalink

https://hdl.handle.net/2142/117730 Copy

Description

Title

Modeling and analysis of trustworthy systems using extensions of network reliability

Author(s)

Nguyen, Hoang Hai

Issue Date

2022-10-25

Director of Research (if dissertation) or Advisor (if thesis)

Nicol, David M

Doctoral Committee Chair(s)

Nicol, David M

Committee Member(s)

• Borisov, Nikita
• Iyer, Ravishankar
• Srikant, Rayadurgam

Department of Study

Electrical & Computer Eng

Discipline

Electrical & Computer Engr

Degree Granting Institution

University of Illinois at Urbana-Champaign

Degree Name

Ph.D.

Degree Level

Dissertation

Keyword(s)

• network security
• uncertain graph
• probabilistic risk assessment
• simulation

Abstract

Modern networked computer systems face an ever-increasing threat of cyber-attacks, whose impacts may range from data loss, financial loss, equipment damage, and in the extreme cases such as safety-critical systems, loss of human lives. Defending such systems is a daunting task, in part due to their complexity and dynamicity and in part due to the uncertainty and constantly evolving nature of cyber-attacks. Facing such a challenge, network defenders need to know how much time, money, and other resources to invest in cyber-defense. Furthermore, given a defense budget, they also need a way to rank the effectiveness of different defense solutions in terms of risk reduction. These questions naturally arise from the domain of risk analysis. However, cybersecurity risk management (CSRM) as a relatively young field lacks sound methods to address the peculiarities and unique challenges of cybersecurity. To that end, this thesis organizes efforts to answer a fundamental question, "how to accurately and efficiently estimate the risk of cyber-attacks against networks?'' As we will see in subsequent chapters, answering this question requires a multidisciplinary approach encompassing network security, graph theory, probabilistic risk assessment, and Monte Carlo simulation. The results of the study are different modeling, analytical, and numerical techniques that can be plugged into a CSRM framework. Together, they help defenders make timely, optimal, and risk-informed decisions in a proactive and reactive fashion--i.e., before and after the event of a network breach. In addition, some techniques can be generalized to a broader class of dependability problems, such as reliability analysis of large-scale systems.

Graduation Semester

2022-12

Type of Resource

Thesis

Copyright and License Information

Copyright 2022 Hoang Hai Nguyen

Owning Collections