In Things We Trust? Towards trustability in the Internet of Things

This essay discusses the main privacy, security and trustability issues with the Internet of Things.

💡 Research Summary

The paper provides a comprehensive overview of the Internet of Things (IoT), beginning with its historical roots in ubiquitous computing and tracing its evolution through RFID, NFC, GPS‑enabled smartphones, and large‑scale services such as Pachube and Flukso. It defines IoT as a dynamic global network where physical and virtual “things” possess identities, attributes, and personalities, are equipped with intelligent interfaces, and are seamlessly integrated into the information network. From this definition the authors extract five core properties: invisibility (devices are embedded in everyday objects and often lack a direct UI), networked (massively distributed, dynamic connections), many‑to‑many relationships (devices are not bound to a single user), always‑on (continuous operation without explicit activation), and context‑awareness (knowledge of location, environment, and nearby devices). These properties enable ambient intelligence but also create novel challenges for privacy, security, and trustability.

Privacy is framed not merely as confidentiality but as a broader data‑protection regime that must respect proportionality and subsidiarity. IoT services inherently require large amounts of personal data to deliver context‑aware functionality, creating a tension between data collection needs and the right to be let alone. Security concerns focus on integrity, authenticity, and availability rather than confidentiality alone. The authors illustrate risks using RFID‑based supply‑chain examples: counterfeit tags, radio jamming, and tag swapping can disrupt inventory management, while compromised actuators in health‑care or smart‑grid contexts could cause physical harm. Moreover, the lack of a clear point of authority in pervasive environments raises questions of accountability.

Trustability is presented as a shift from the traditional “trust‑us” model to a user‑centric paradigm where individuals can assess risk using third‑party tools and data under their own control. The paper proposes a definition: a system is trustable if a user can reliably estimate the risk of using it for a specific purpose through independently controlled tools or data sources. This perspective suggests re‑applying identity‑management and trusted‑computing concepts to IoT.

Implementing privacy and security guarantees must respect the constraints of IoT devices: limited power, minimal computational resources, massive scale, and the aforementioned properties. Classical security mechanisms are often too heavyweight for low‑cost RFID tags, which cannot perform more than primitive cryptographic operations.

The authors review prior work on data minimisation, which includes tag identifier relabelling, periodic re‑encryption, hash‑chain based identifier updates, and key‑tree structures for symmetric key distribution. While these schemes reduce traceability and improve forward privacy, they suffer from synchronization issues, key‑search overhead, and impracticality when millions of tags each hold unique secret keys.

Recent alternative approaches aim to restore user control. Examples such as the RFID Guardian, Privacy Coach, and the “resurrecting duckling” principle provide mechanisms for users to block or selectively enable tag reads, enforce fine‑grained access policies, and manage trust relationships. These tools act as complementary layers rather than complete replacements for cryptographic protection.

In conclusion, the paper argues that a holistic, multi‑layered framework is required for the IoT: lightweight cryptography, policy‑based access control, transparent accountability, and user‑driven trust‑building mechanisms must be combined. Open research challenges include standardised trustability metrics, real‑time privacy management, scalable key and credential management for massive deployments, and legal‑policy frameworks that define responsibility in pervasive environments. The work calls for interdisciplinary efforts to ensure that the IoT evolves into a safe, open, and trustworthy ecosystem.