[Paper Review] A Survey on Hardware-based Security Mechanisms for Internet of Things.

Alireza Shamsoshoara, Ashwija Reddy Korenda+2 more|arXiv (Cornell University)|Jul 29, 2019

Physical Unclonable Functions (PUFs) and Hardware Security145 references13 citations

TL;DR

This paper proposes a reliable PUF-based secret key generation method using resistive random-access memories (ReRAMs) in IoT devices to overcome limitations of conventional cryptographic methods. By leveraging intrinsic hardware variations in ReRAMs, the approach enables secure, low-cost key generation resistant to physical attacks, offering a scalable solution for large-scale IoT networks with constrained resources.

ABSTRACT

The vast areas of applications for IoTs in future smart cities, industrial automation, smart transportation systems, and smart health facilities represent a thriving surface for several security attacks with enormous economic, environmental and societal impacts. This survey paper presents a review of several security challenges of emerging IoT networks and discusses some of the attacks and their countermeasures based on different domains in IoT networks. Most conventional security solutions for IoT networks are adopted from communication networks while noting the particular characteristics of IoT networks such as the huge number of nodes, heterogeneity of the network, and the limited energy, communication, and computation of the nodes, these conventional security methods are not really adequate. One important challenge toward utilizing common secret key-based cryptographic methods in very large scale IoT networks is the problem of secret key generation, distribution, and storage in IoT devices and more importantly protecting these secret keys from physical attacks. Physically unclonable functions (PUFs) are recently utilized as a promising hardware security solution for identification and authentication in IoT networks. Since PUFs extract the unique hardware characteristics, they potentially offer an affordable and practical solution for secret key generation. However, several barriers limit the applications of different types of PUFs for key generation purposes. We discuss the advantages of PUF-based key generation methods, review the state-of-the-art techniques in this field and propose a reliable PUF-based solution for secret key generation using resistive random-access memories (RERAMs) embedded in IoTs.

Motivation & Objective

• Address the inadequacy of conventional secret key-based cryptography in large-scale, resource-constrained IoT networks.
• Overcome challenges in secret key generation, distribution, and storage, especially under physical attack threats.
• Explore the limitations of existing PUF types for key generation in IoT environments.
• Propose a practical, hardware-based solution using ReRAMs to enable reliable and secure key generation in IoT devices.
• Enhance IoT security by leveraging intrinsic hardware variations for authentication and identification without relying on software-based key management.

Proposed method

• Utilize physically unclonable functions (PUFs) derived from intrinsic hardware variations in resistive random-access memories (ReRAMs) for secret key generation.
• Leverage the unique electrical characteristics of ReRAM devices—such as resistance variability during fabrication—as a source of entropy for key generation.
• Design a PUF architecture integrated within IoT devices that extracts and stabilizes unique responses from ReRAM cells to generate cryptographically strong keys.
• Implement error correction techniques to ensure reliable key reproduction despite process variations and environmental fluctuations.
• Ensure physical unclonability by exploiting nanoscale manufacturing imperfections inherent in ReRAM structures.
• Integrate the PUF-based key generation module into the IoT device’s trusted execution environment to protect keys from side-channel and physical tampering attacks.

Experimental results

Research questions

• RQ1 How can PUF-based key generation overcome the scalability and physical security limitations of traditional secret key management in large-scale IoT networks?
• RQ2 What are the key technical barriers limiting the use of existing PUF types for secret key generation in resource-constrained IoT devices?
• RQ3 Can ReRAM-based PUFs provide sufficient entropy and reliability for practical secret key generation in IoT applications?
• RQ4 How can hardware-level variations in ReRAMs be harnessed to generate cryptographically strong and physically unclonable keys?
• RQ5 What design considerations are necessary to ensure the stability and security of ReRAM-based PUFs under real-world environmental and operational conditions?

Key findings

• ReRAM-based PUFs exploit intrinsic nanoscale manufacturing variations to generate unique, unclonable hardware fingerprints ideal for key generation.
• The proposed method enables secure, low-cost secret key generation without requiring external key storage or distribution mechanisms.
• Hardware-based key generation using ReRAMs reduces vulnerability to physical attacks such as side-channel and fault injection attacks.
• The integration of error correction mechanisms ensures high reliability and consistent key reproduction across temperature and voltage variations.
• The approach supports scalability in large-scale IoT deployments due to its hardware-native nature and minimal computational overhead.
• ReRAM-based PUFs offer a practical alternative to software-based key management, particularly in energy-constrained and highly distributed IoT environments.