Towards building trustworthy blockchain-enabled decentralized applications: Data feeds reliability, multi-service authentication, and sustainable crop insurance
Zhonghao Liao · 2024
Blockchain has been recognized as one of the most transformative technologies poised to revolutionize the digital infrastructure landscape, with applications spanning various industries, including supply chain management, financial services, smart cities, and precision agriculture. Its distributed and tamper-resistant nature offers promising solutions to many problems inherent in centralized systems. However, blockchain-enabled applications face significant challenges, such as inefficient data storage, privacy concerns, reliable data input, and user accessibility issues. These challenges hinder the wider adoption of blockchain technology and limit its potential to fully transform and enhance various sectors. Addressing these obstacles is crucial for unlocking the full capabilities of blockchain and realizing its promise of a more secure, efficient, and transparent digital future. This dissertation addresses these limitations through a series of innovative solutions. We present AgBIS, the first blockchain-enabled crop insurance platform specifically designed for Multi-peril Crop Insurance (MPCI) programs. AgBIS demonstrates the potential of blockchain in financial services, particularly in reducing information asymmetry, thereby mitigating issues such as adverse selection, moral hazard, and insurance fraud within crop insurance ecosystems. In the meantime, AgBIS tackles inefficiencies and ensures the confidentiality of on-chain data storage. To address the issue of reliable data feeds, we present TAHOE, a trustworthy hardware oracle scheme. TAHOE leverages physical unclonable functions (PUFs) to ensure the authenticity and integrity of the data fed into the blockchain network. This work also explores the Discharge Inversion Effect (DIE) in collecting power-up states of SRAM chips, which impacts the reliability of PUF-based cryptographic applications unless proper guidelines are followed. We further propose OracDiFT to enhance data quality assurance and validity without direct access to the original data collection process. By providing on-demand data quality assurance, OracDiFT ensures high-quality data from collection to final use, which is crucial for robust data collection, utilization, and sharing in applications such as financial services and healthcare domains. Lastly, the usability and security of blockchain-enabled services are improved with RUDBA, a reusable user-device biometric authentication scheme. RUDBA combines biometric information with the device's digital fingerprint, providing a dual-layer security mechanism that leads to a reliable and reusable user-device authentication system in multi-service systems.