A Lightweight Protocol for Secure Data Provenance in the Internet of Things Using Wireless Fingerprints

Abstract

Data provenance is an important security requirement to establish trust in the data produced by an Internet of Things (IoT) device. Existing works on data provenance for IoT are based on complex computations or costly hardware that may not be feasible for IoT systems. To solve this issue, this article uses an analytical model to develop a threshold-based mechanism to establish data provenance in IoT systems. Moreover, using lightweight security primitives, a lightweight security protocol for data provenance is also proposed. The proposed protocol uses physical unclonable functions and fingerprints extracted from the wireless channel to achieve data provenance, mutual authentication, and anonymity. The wireless fingerprints are generated using the link quality indicator values. Experimental validation on MICA Z motes shows that the proposed technique can detect adversarial channels with high accuracy. Security analysis of the proposed protocol using formal proofs as well as simulations shows the robustness against various types of attacks. Moreover, the energy requirements for the proposed protocol are shown to be significantly lower than existing protocols.