Please use this identifier to cite or link to this item: https://hdl.handle.net/2440/137544
Type: Thesis
Title: Constructing Security Functions for Resource-Limited Devices With Memory Fingerprints
Author: Su, Yang
Issue Date: 2023
School/Discipline: School of Computer and Mathematical Science
Abstract: The exponential rate of hardware miniaturisation, the emergence of low cost and low power sensing modalities coupled with rapid developments in communication technologies are driving the world towards a future where tiny scale computing will be more pervasive and seamlessly integrated with everyday life. This sea of change is driven by the increasing ability of tiny computing platforms to connect people and things to the Internet—the Internet-of-Things (IoT)—enabling transformative applications ranging from connected healthcare to smart cities. More recently, we have seen the emergence of such tiny scale computing platforms in the form of highly resource-constrained and intermittently powered batteryless devices that rely only on harvested RF (radio frequency) energy for operations, best exemplified by Computational Radio Frequency Identification (CRFID) devices. Despite the simplicity of deployment, reduction in cost and perpetual operational life offered by such devices, provision of security services, offered to typical computing platforms, is significantly more challenging for CRFID devices due to: intermittent powering, unavailability of hardware security support, constrained air interface protocols, lack of secure storage space, limited computational capabilities, and the absence of a supervisory operating system. This dissertation focuses on addressing the challenging problems associated with the provision of security services to resource-limited and intermittently powered devices exemplified by CRFID technologies. Given the need to update the firmware of such devices, the thesis investigates how a secure and wireless code update mechanism that is compliant with current communication protocols can be realised under resource constraints and intermittent powering without additional hardware components. The thesis presents a rigorous design, development and implementation of the first secure wireless firmware update scheme for CRFID devices based on entangling a volatile and hardware instance specific secret key from the on-chip SRAM to the firmware update mechanism. The method, called SecuCode, only allows an authorised party to perform a wireless firmware update and does not require any hardware modifications whilst being standards-compliant. The update methods are further developed for simultaneous and secure wireless firmware update of multiple CRFID devices to prevent security threats such as malicious code injection, IP theft, and incomplete code installation whilst complying with standard hardware and protocols. The proposed method, called Wisecr, facilitate a secure and scalable method of code update for battery-free passively powered CRFID devices. Given the lack of secure storage and the prohibitive cost of providing such storage under cost, power and resource constraints, the thesis investigates exploiting ubiquitously available memory fingerprints for security functions on resource-limited devices. Device memory fingerprints generated at different time instances are susceptible to unpredictable noise. The state-of-the-art reverse fuzzy extractor (RFE)-based method has been demonstrated to derive usable keys from the inherently unreliable device fingerprints for security functions. However, the computationally-intensive nature of the on-device RFE encoder renders it challenging to employ RFE-based methods on resource-constrained devices. The thesis first proposes a multiple referenced responses (MRR) strategy for device fingerprint enrolment. The proposed approach significantly reduces the on-device implementation overheads for the RFE encoder. The thesis then investigates the transformation of raw memory fingerprints into a noise-tolerant space where the generated device fingerprints are intrinsically highly reliable. The proposed method, NoisFre, fundamentally removes the need for an RFE encoder from reliable device fingerprint key derivation methods. The thesis investigates and proposes NoisFre-Lite to further improve the extraction efficiency of the noise-tolerant fingerprints to enable mounting NoisFre on devices with limited memory sizes. To this end, a highly reliable yet lightweight key generation from ubiquitously available memory fingerprints is achieved for devices with computation and resource limitations to realise the practical use of memory fingerprints for security functions.
Advisor: Ranasinghe, Damith Chinthana
Kavehei, Omid
Dissertation Note: Thesis (Ph.D.) -- University of Adelaide, School of Computer and Mathematical Science, 2023
Keywords: Internet of Things, Secure Wireless Firmware Update, Secret Key Generation, Secret Key Management, Resource-Constrained Devices, Computational Radio Frequency Identification, Physically Unclonable Functions, Device Memory Fingerprinting, Intermittent Powering
Provenance: This electronic version is made publicly available by the University of Adelaide in accordance with its open access policy for student theses. Copyright in this thesis remains with the author. This thesis may incorporate third party material which has been used by the author pursuant to Fair Dealing exceptions. If you are the owner of any included third party copyright material you wish to be removed from this electronic version, please complete the take down form located at: http://www.adelaide.edu.au/legals
Appears in Collections:Research Theses

Files in This Item:
File Description SizeFormat 
Su2023_PhD.pdf51.78 MBAdobe PDFView/Open


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.