DOI
https://doi.org/10.25772/F6EQ-7609
Defense Date
2022
Document Type
Dissertation
Degree Name
Doctor of Philosophy
Department
Computer Science
First Advisor
Dr.Tamer Nadeem
Abstract
The rapid growth of edge-based IoT devices, their use cases, and autonomous communication has created new challenges with privacy and security. Side-channel attacks are one of the examples of security and privacy vulnerabilities that can cause inference at Internet-Service Provider (ISP) and local Wi-Fi networks. Such an attack would leak user’s sensitive information such as home occupancy, medical activity, and daily routines. Another example is that these devices have weak authentication and low encryption standards, making them an easy target for malware-based attacks such as denial of service or launching other network attacks using these infected devices. This thesis dissertation explored ML-assisted tools to secure the devices from network-based security and privacy inference attacks. In this thesis, one of our components, securing edge devices against malware-based attacks or anomalies, focused on building an ML-assisted security service and evaluating it for different real-world conditions such as robustness to noise, adaptability to new devices, and less resource-intensive deployment. Moreover, given the recent trend of adversarial attacks against machine learning models, we designed a novel edge-based adversarial attack against edge-based machine learning models and explored techniques to make edge-based network security services robust. Another component of this thesis focused on building a privacy-preserving service to obfuscate and minimize the privacy inference on network communication. The main objective of this component was to evaluate different privacy-preserving techniques such as traffic shaping and injecting synthetic network traffic for countering the effectiveness of side-channel attacks.
Rights
© The Author
Is Part Of
VCU University Archives
Is Part Of
VCU Theses and Dissertations
Date of Submission
5-11-2022