Date of Award

Fall 2019

Document Type


Degree Name

Doctor of Philosophy (PhD)


Electrical & Computer Engineering

Committee Director

Sachin Shetty

Committee Member

Chunsheng Xin

Committee Member

Dimitrie Popescu

Committee Member

Rao Chaganty

Committee Member

Laurent Njilla


Stepping stone attacks are one of the most sophisticated cyber-attacks, in which attackers make a chain of compromised hosts to reach a victim target. In this Dissertation, an analytic model with Multi-Agent systems approach has been proposed to analyze the propagation of stepping stones attacks in dynamic vulnerability graphs. Because the vulnerability configuration in a network is inherently dynamic, in this Dissertation a biased min-consensus technique for dynamic graphs with fixed and switching topology is proposed as a distributed technique to calculate the most vulnerable path for stepping stones attacks in dynamic vulnerability graphs. We use min-plus algebra to analyze and provide necessary and sufficient convergence conditions to the shortest path in the fixed topology case. A necessary condition for the switching topology case is provided.

Most cyber-attacks involve an attacker launching a multi-stage attack by exploiting a sequence of hosts. This multi-stage attack generates a chain of ``stepping stones” from the origin to target. The choice of stepping stones is a function of the degree of exploitability, the impact, attacker’s capability, masking origin location, and intent. In this Dissertation, we model and analyze scenarios wherein an attacker employs multiple strategies to choose stepping stones. The problem is modeled as an Adjacency Quadratic Shortest Path using dynamic vulnerability graphs with multi-agent dynamic system approach. With this approach, the shortest stepping stone path with maximum node degree and the shortest stepping stone path with maximum impact are modeled and analyzed.

Because embedded controllers are omnipresent in networks, in this Dissertation as a Risk Mitigation Strategy, a cyber-attack tolerant control strategy for embedded controllers is proposed. A dual redundant control architecture that combines two identical controllers that are switched periodically between active and restart modes is proposed. The strategy is addressed to mitigate the impact due to the corruption of the controller software by an adversary. We analyze the impact of the resetting and restarting the controller software and performance of the switching process. The minimum requirements in the control design, for effective mitigation of cyber-attacks to the control software that implies a “fast” switching period is provided. The simulation results demonstrate the effectiveness of the proposed strategy when the time to fully reset and restart the controller is faster than the time taken by an adversary to compromise the controller. The results also provide insights into the stability and safety regions and the factors that determine the effectiveness of the proposed strategy.


In Copyright. URI: This Item is protected by copyright and/or related rights. You are free to use this Item in any way that is permitted by the copyright and related rights legislation that applies to your use. For other uses you need to obtain permission from the rights-holder(s).