Enhancing Security in Internet of Things (IoT) Systems through Data Provenance

Abstract

The Internet of Things (IoT) is simply a collection of billions of resource-limited, device heterogeneity and constantly producing/delivering data on untrusted networks that perimeter-based security mechanisms are unable to defend. Many IoT security incidents are not because of a single point of failure, but because the data source is unknown, the data is altered, or whether it is trusted or not before it makes a decision. In this paper, the authors explore the concept of data provenance, defined as the systematic documentation of the origin, custody and evolution of data, as a key mechanism to improve IoT security. The paper examines the threat landscape of IoT and surveys various approaches to provenance capture and verification, such as blockchain-based ledgers, hardware-rooted trust anchors, cryptographic hash chaining, and machine-learning assisted provenance graph analysis, and introduces a five-layer provenance-enabled security architecture that combines lightweight cryptographic tagging at the perception layer, with distributed ledger commitment and anomaly-aware auditing at the application layer. Two summary tables provide a comparative analysis of the techniques of provenance on overhead, tamper-resistance and scalability axes, and three explanatory figures help clarify the concepts. The benefits of security guarantees are discussed along with the restricted computational, memory and energy resources of IoT devices; open challenges such as standardisation, interoperability, proving it is not quantum-resistant, and disclosure with privacy are identified. Overall, the paper suggests that provenance-aware security, in conjunction with lightweight cryptography and edge-based verification, provides a viable and scalable approach to secure and trustworthy IoT ecosystems.