Over the last two decades, hardware security has gained increasing attention in academia and industry. Flash memory has been given a spotlight in recent years, with the question of whether or not it can prove useful in a security role. Because of inherent process variation in the characteristics of flash memory modules, they can provide a unique fingerprint for a device and have thus been proposed as locations for hardware security primitives. These primitives include physical unclonable functions (PUFs), true random number generators (TRNGs), and integrated circuit (IC) counterfeit detection. In this paper, we evaluate the efficacy of flash memory-based security primitives and categorize them based on the process variations they exploit, as well as other features. We also compare and evaluate flash-based security primitives in order to identify drawbacks and essential design considerations. Finally, we describe new directions, challenges of research, and possible security vulnerabilities for flash-based security primitives that we believe would benefit from further exploration.

No abstract

IoT devices have become popular targets for various network attacks due to their lack of industry-wide security standards. In this work, we focus on smart home IoT device identification and defending them against Distributed Denial of Service (DDoS) attacks. The proposed framework protects smart homes by using VLAN-based network isolation. This architecture has two VLANs: one with non-verified devices and the other with verified devices, both of which are managed by the SDN controller. Lightweight stateless flow-based features, including ICMP, TCP, and UDP protocol percentage, packet count and size, and IP diversity ratio, are proposed for efficient feature collections. Further analysis is performed to minimize training data to run on resource-constrained edge devices in smart home networks. Three popular machine learning algorithms, including K-Nearest-Neighbors, Random Forest, and Support Vector Machines, are used to classify IoT devices and detect different types of DDoS attacks, including TCP-SYN, UDP, and ICMP. The system's effectiveness and efficiency are evaluated by emulating a network consisting of an Open vSwitch, Faucet SDN controller, and several IoT device traces from two different testbeds.

No abstract

Since the inception of encrypted messages thousands of years ago, mathematicians and scientists have continued to improve encryption algorithms in order to create more secure means of communication. These improvements came by means of more complex encryption algorithms that have stronger security features such as larger keys and trusted third parties. While many new processors can handle these more complex encryption algorithms, IoT devices on the edge often struggle to keep up with resource intensive encryption standards. In order to meet this demand for lightweight, secure encryption on the edge, this paper proposes a novel solution, called the High and Low (HaLo) method, that generates Physical Unclonable Function (PUF) signatures based on process variations within flash memory. These PUF signatures can be used to uniquely identify and authenticate remote sensors, and help ensure that messages being sent from remote sensors are encrypted adequately without requiring computationally expensive methods. The HaLo method consumes 20x less power than conventional authentication schemes commonly used with IoT devices, it has an average latency of only 39ms for 512 bit signature generation, and the average error rate is below 0.06%. Due to its low latency, low error rate, and high power efficiency, the HaLo method can progress the field of IoT encryption standards by accurately and efficiently authenticating remote sensors without sacrificing encryption integrity.