Date on Master's Thesis/Doctoral Dissertation

8-2021

Document Type

Doctoral Dissertation

Degree Name

Ph. D.

Department

Computer Engineering and Computer Science

Degree Program

Computer Science and Engineering, PhD

Committee Chair

Lauf, Adrian

Committee Member

Elmaghraby, Adel

Committee Member

Kantardzic, Mehmed

Committee Member

Lovavio, Michael

Author's Keywords

Security; internet of things; industrial IoT; internet of vehicles; authentication; smart homes; smart city

Abstract

Internet of Things (IoT) presents a holistic and transformative approach for providing services in different domains. IoT creates an atmosphere of interaction between humans and the surrounding physical world through various technologies such as sensors, actuators, and the cloud. Theoretically, when everything is connected, everything is at risk. The rapid growth of IoT with the heterogeneous devices that are connected to the Internet generates new challenges in protecting and preserving user’s privacy and ensuring the security of our lives. IoT systems face considerable challenges in deploying robust authentication protocols because some of the IoT devices are resource-constrained with limited computation and storage capabilities to implement the currently available authentication mechanism that employs computationally expensive functions. The limited capabilities of IoT devices raise significant security and privacy concerns, such as ensuring personal information confidentiality and integrity and establishing end-to-end authentication and secret key generation between the communicating device to guarantee secure communication among the communicating devices. The ubiquity nature of the IoT device provides adversaries more attack surfaces which can lead to tragic consequences that can negatively impact our everyday connected lives. According to [1], authentication and privacy protection are essential security requirements. Therefore, there is a critical need to address these rising security and privacy concerns to ensure IoT systems' safety. This dissertation identifies gaps in the literature and presents new mutual authentication and privacy preservation schemes that fit the needs of resource-constrained devices to improve IoT security and privacy against common attacks. This research enhances IoT security and privacy by introducing lightweight mutual authentication and privacy preservation schemes for IoT based on hardware biometrics using PUF, Chained hash PUF, dynamic identities, and user’s static and continuous biometrics. The communicating parties can anonymously communicate and mutually authenticate each other and locally establish a session key using dynamic identities to ensure the user’s unlinkability and untraceability. Furthermore, virtual domain segregation is implemented to apply security policies between nodes. The chained-hash PUF mechanism technique is implemented as a way to verify the sender’s identity. At first, this dissertation presents a framework called “A Lightweight Mutual Authentication and Privacy-Preservation framework for IoT Systems” and this framework is considered the foundation of all presented schemes. The proposed framework integrates software and hardware-based security approaches that satisfy the NIST IoT security requirements for data protection and device identification. Also, this dissertation presents an architecture called “PUF Hierarchal Distributed Architecture” (PHDA), which is used to perform the device name resolution. Based on the proposed framework and PUF architecture, three lightweight privacy-preserving and mutual authentication schemes are presented. The Three different schemes are introduced to accommodate both stationary and mobile IoT devices as well as local and distributed nodes. The first scheme is designed for the smart homes domain, where the IoT devices are stationary, and the controller node is local. In this scheme, there is direct communication between the IoT nodes and the controller node. Establishing mutual authentication does not require the cloud service's involvement to reduce the system latency and offload the cloud traffic. The second scheme is designed for the industrial IoT domain and used smart poultry farms as a use case of the Industrial IoT (IIoT) domain. In the second scheme, the IoT devices are stationary, and the controller nodes are hierarchical and distributed, supported by machine-to-machine (M2M) communication. The third scheme is designed for smart cities and used IoV fleet vehicles as a use case of the smart cities domain. During the roaming service, the mutual authentication process between a vehicle and the distributed controller nodes represented by the Roadside Units (RSUs) is completed through the cloud service that stores all vehicle's security credentials. After that, when a vehicle moves to the proximity of a new RSU under the same administrative authority of the most recently visited RSU, the two RSUs can cooperate to verify the vehicle's legitimacy. Also, the third scheme supports driver static and continuous authentication as a driver monitoring system for the sake of both road and driver safety. The security of the proposed schemes is evaluated and simulated using two different methods: security analysis and performance analysis. The security analysis is implemented through formal security analysis and informal security analysis. The formal analysis uses the Burrows–Abadi–Needham logic (BAN) and model-checking using the automated validation of Internet security protocols and applications (AVISPA) toolkit. The informal security analysis is completed by: (1) investigating the robustness of the proposed schemes against the well-known security attacks and analyze its satisfaction with the main security properties; and (2) comparing the proposed schemes with the other existing authentication schemes considering their resistance to the well-known attacks and their satisfaction with the main security requirements. Both the formal and informal security analyses complement each other. The performance evaluation is conducted by analyzing and comparing the overhead and efficiency of the proposed schemes with other related schemes from the literature. The results showed that the proposed schemes achieve all security goals and, simultaneously, efficiently and satisfy the needs of the resource-constrained IoT devices.

Share

COinS