Date of Award

Summer 8-2022

Document Type


Degree Name

Doctor of Philosophy (PhD)


Computational Modeling & Simulation Engineering


Modeling and Simulation

Committee Director

Sachin Shetty

Committee Member

Yuzhong Shen

Committee Member

Hong Yang

Committee Member

Chunsheng Xin


The high connectivity of modern cyber networks and devices has brought many improvements to the functionality and efficiency of networked systems. Unfortunately, these benefits have come with many new entry points for attackers, making systems much more vulnerable to intrusions. Thus, it is critically important to protect cyber infrastructure against cyber attacks. The static nature of cyber infrastructure leads to adversaries performing reconnaissance activities and identifying potential threats. Threats related to software vulnerabilities can be mitigated upon discovering a vulnerability and-, developing and releasing a patch to remove the vulnerability. Unfortunately, the period between discovering a vulnerability and applying a patch is long, often lasting five months or more. These delays pose significant risks to the organization while many cyber networks are operational. This concern necessitates the development of an active defense system capable of thwarting cyber reconnaissance missions and mitigating the progression of the attacker through the network. Thus, my research investigates how to develop an efficient defense system to address these challenges. First, we proposed the framework to show how the defender can use the network of decoys along with the real network to introduce mistrust. However, another research problem, the defender’s choice of whether to save resources or spend more (number of decoys) resources in a resource-constrained system, needs to be addressed. We developed a Dynamic Deception System (DDS) that can assess various attacker types based on the attacker’s knowledge, aggression, and stealthiness level to decide whether the defender should spend or save resources. In our DDS, we leveraged Software Defined Networking (SDN) to differentiate the malicious traffic from the benign traffic to deter the cyber reconnaissance mission and redirect malicious traffic to the deception server. Experiments conducted on the prototype implementation of our DDS confirmed that the defender could decide whether to spend or save resources based on the attacker types and thwarted cyber reconnaissance mission. Next, we addressed the challenge of efficiently placing network decoys by predicting the most likely attack path in Multi-Stage Attacks (MSAs). MSAs are cyber security threats where the attack campaign is performed through several attack stages and adversarial lateral movement is one of the critical stages. Adversaries can laterally move into the network without raising an alert. To prevent lateral movement, we proposed an approach that combines reactive (graph analysis) and proactive (cyber deception technology) defense. The proposed approach is realized through two phases. The first phase predicts the most likely attack path based on Intrusion Detection System (IDS) alerts and network trace. The second phase determines the optimal deployment of decoy nodes along the predicted path. We employ transition probabilities in a Hidden Markov Model to predict the path. In the second phase, we utilize the predicted attack path to deploy decoy nodes. The evaluation results show that our approach can predict the most likely attack paths and thwart adversarial lateral movement.