Fish, Robert S.Eze, Okezie2025-08-062025-08-062025-04-11https://theses-dissertations.princeton.edu/handle/88435/dsp01wm117s43zInternet of Things (IoT) devices promise transformative benefits across industries, yet security weaknesses—particularly in the firmware-update process—pose critical challenges. This thesis examines whether a decentralized, blockchain-based update framework can enhance IoT security without imposing unacceptable performance costs. A controlled IoT testbed of Arduino-based sensor nodes was constructed to empirically compare three network architectures for remote firmware updates: a local-server model, a cloud-server model, and a blockchain-integrated model built on a private Ethereum network. Each architecture’s firmware-update performance—latency for discovery, download, and installation—was measured, and the resilience of each design to cyberattacks was evaluated under simulated Distributed Denial-of-Service (DDoS) and Man-in-the-Middle (MitM) scenarios. Experimental results show that although the two traditional client–server models deliver updates with lower latency and higher efficiency, the blockchain-based system markedly improves security, preventing unauthorized or malicious updates and remaining robust against attacks that compromised the centralized models. This security–performance trade-off indicates that blockchain technology, despite modest additional latency, can substantially harden IoT networks against firmware tampering. In safety-critical IoT deployments, the enhanced trustworthiness of a decentralized update mechanism may outweigh its efficiency costs, demonstrating that blockchain is a viable and valuable solution for secure IoT device maintenance.en-USPrinceton University Senior Theses