Survey of the state-of-the-art data provenance approaches and their application in IoT platforms
What is it about?
Since building a completely secure system is impossible, a security-critical system needs to be resilient against cyberattacks, and data provenance can help achieve that. Data provenance provides us with a comprehensive history of data and processes across a system. We can leverage this history to verify data ownership and track the process activities and the changes made in data. We can also employ this history in reconstructing an attack scenario to understand the attack behavior better. In this paper, we survey the state-of-the-art provenance approaches that exist in the literature and present a comprehensive comparative discussion of those approaches. But before that, we provide a conceptual overview of a provenance system architecture. We also discuss how we can deploy provenance systems in an IoT environment to ensure policy compliance, detect attack chains in trigger-action platforms, and maintain data integrity. Since commodity and health IoT devices continuously collect user-sensitive data, we need to verify whether the devices comply with the defined policies and whether there are vulnerabilities in trigger-action platforms that attackers can exploit to attack the network and steal user data. In addition, we also present a brief discussion on how to secure a provenance system itself so that attackers cannot alter the components of the provenance system before attempting the ultimate attack. Finally, we recommend some probable research topics that may enrich the existing literature.
Why is it important?
We present a comprehensive overview of the recent provenance approaches, including those which support on-the-fly provenance reduction and runtime provenance analysis. We also provide a comparative discussion on the state-of-the-art provenance systems used in the IoT domain and briefly discuss the security of provenance systems themselves. Thus, this survey paper gives a complete overview of the existing provenance approaches and demonstrates how these approaches can be securely employed to ensure the security and privacy of IoT and other distributed platforms. Furthermore, our suggestion of future research directions will help the provenance community advance the literature to a certain extent.