Understanding the Security Landscape of Embedded Non-Volatile Memories: A Comprehensive Survey

Zakia Tamanna Tisha, Ujjwal Guin · 2025

Read Understanding the Security Landscape of Embedded Non-Volatile Memories: A Comprehensive Survey by Zakia Tamanna Tisha, Ujjwal Guin (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2505-17253.txt.

Introduction

Zakia Tamanna Tisha and Ujjwal Guin survey the security of embedded non-volatile memories, in a paper from 2025. These memories keep data without power and offer fast access in small, low-energy packages. The same trait that makes them useful, holding data through power loss, also opens security holes.

The survey explains, from the design level, why specific memory types are exposed. It reviews security tools built from these memories, such as fingerprint-like hardware keys and random number generators, plus methods that hide circuit logic. Threats range from physical attacks, like probing and fault injection, to logical ones, like data leakage and denial of service. The paper maps the field for designers who must weigh speed against safety.

Catalog record

CREATORZakia Tamanna Tisha, Ujjwal Guin
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE62 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2505-17253

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Understanding the Security Landscape of Embedded Non-Volatile Memories: A Comprehensive Survey Zakia Tamanna Tisha, Ujjwal Guin Abstract The modern semiconductor industry requires memory solutions that can keep pace with the high-speed demands of high-performance computing. Embedded non-volatile memories (eNVMs) address these requirements by offering faster access to stored data at an improved computational throughput and efficiency. Furthermore, these technologies offer numerous appealing features, including limited area-energy-runtime budget and data retention capabilities. Among these, the data retention feature of eNVMs has garnered particular interest within the semiconductor community. Although this property allows eNVMs to retain data even in the absence of a continuous power supply, it also introduces some vulnerabilities, prompting security concerns. These concerns have sparked increased interest in examining the broader security implications associated with eNVM technologies. This paper examines the security aspects of eNVMs by discussing the reasons for vulnerabilities in specific memories from an architectural point of view. Additionally, this paper extensively reviews eNVM-based security primitives, such as physically unclonable functions and true random number generators, as well as techniques like logic obfuscation. The paper also explores a broad spectrum of security threats to eNVMs, including physical attacks such as side-channel attacks, fault injection, and probing, as well as logical threats like information leakage, denial-of-service, and thermal attacks. Finally, the paper presents a study of publication trends in the eNVM domain since the early 2000s, reflecting the rising momentum and research activity in this field. K eywords Non-volatile memories \cdot security primitives \cdot PUFs \cdot TRNGs \cdot side-channel analysis \cdot probing \cdot fault injection 1 Introduction The demand for memory devices has evolved significantly over the past few decades. This shift is primarily driven by factors such as scaling and the demand for high-density and faster memory devices. Traditional memory technologies like SRAM and DRAM have been the building blocks of computing systems for years. Although they provide fast memory access, their biggest limitation lies in their dependence on continuous power to retain information. Flash memory revolutionized storage by enabling mobility applications. However, as semiconductor scaling reaches its physical limits, flash memory struggles to meet the market demand for more endurance, write speed, and power efficiency. These limitations have set the stage for next-generation eNVM technologies, offering significant advantages such as non-volatility and enhanced performance. With their advanced architectures, eNVMs are becoming essential to modern high-performance computing systems. Beyond performance enhancements, eNVMs also serve as integral components in secure hardware systems. They facilitate the design of secure architectures that resist tampering and are used in a variety of applications, including cryptography and secure boot processes. They support scalability, reconfigurability, and cost-effective local computing. They also serve as a rich source of entropy, making them ideal candidates for building security primitives like physically unclonable functions (PUFs) and true random number generators (TRNGs). Nevertheless, the integration of eNVMs into modern computing systems also introduces new risks, such as the potential leakage of sensitive data like secret keys, login credentials, and credit card information [ 1 ] . Historically, data security in volatile memories (like SRAMs and DRAMs) was not a major concern since they lose their content when powered down. Yet, as these cache memories evolve to potentially become non-volatile, there is an increased risk of adversaries accessing sensitive information in its unencrypted form. As the memory landscape continues to evolve, a comprehensive understanding of both aspects of eNVM security has become crucial to developing reliable computing environments. Devices used for security purposes increasingly rely on the unique properties of eNVMs. Their non-volatility enables critical data to persist across power cycles, ensuring secure operation even during power outages. eNVMs also offer tamper resistance for protecting sensitive data like cryptographic keys from physical and logical attacks. Additionally, eNVMs have increased endurance to support frequent updates and record security data, and their fast access speeds help retrieve important information quickly. They also include advanced error correction to maintain data integrity and prevent corruption. In this paper, we dive deep into the security aspects of eNVMs. Our study begins with an in-depth look at various eNVM technologies, exploring the underlying technology behind security devices and the solutions that leverage eNVMs. As the applications of eNVMs grow, so do the risks associated with them. A key part of our research also involves exploring the structural characteristics of eNVMs that make them susceptible to security attacks. Compared to the previous work published in ISVLSI [ 2 ] , we make the following modifications in this paper: • Expanding Vulnerability Study: This study builds on our preliminary research, which explored five different non-volatile memory (NVM) technologies. Here, we provide a more detailed analysis of their security vulnerabilities. In particular, we investigate the architectural configurations and critical parametric choices that make eNVMs susceptible to various security attacks. For example, factors such as array topology, write current, material properties, etc., are shown to impact eNVM security significantly. • Broadening Security Primitive Study: In this work, we explore the technologies underlying security devices and solutions that leverage eNVMs. Building upon the extensive literature reviewed in the preliminary study, additional publications are examined to advance the collective understanding of the field further. The findings indicate that the unique physical and electrical properties of eNVMs enable the development of various security primitives, including PUFs, TRNGs, and logic obfuscation techniques. • Expanding Attack Vectors: We investigate additional security attacks targeting eNVMs that were not addressed in our initial work. These newly examined attacks include information leakage attacks, which exploit side channels or remanence to expose sensitive data; denial of service (DoS) attacks, which aim to disrupt memory operations and impact system availability; and thermal attacks, where temperature manipulation is used to trigger faults or change memory behavior. Examining these security threats provides a broader perspective of the security challenges associated with eNVM-based systems. • Analysis of Research Trends: This study presents a comprehensive analysis of publication trends in the NVM domain since the early 2000s, highlighting the evolution of research focus over time. There has been a significant surge in research activity across diverse NVM technologies, driven by rising interest and accelerated progress in the field. In parallel, we examine the key technological milestones that illustrate how NVM technologies have evolved over the years to address challenges such as energy efficiency, performance scalability, and computing efficiency. By capturing both research trends and innovation trajectories, the study offers valuable insights into the evolution of NVM technologies, helping to shape future directions and innovations in memory design, particularly in the security domain. The rest of the paper is structured as follows: Section 2 covers various NVM technologies and discusses their architectural vulnerabilities. Section 3 delves into security applications that utilize eNVMs. Section 4 examines a broad spectrum of attack vectors targeting eNVMs. Section 5 presents a detailed analysis of NVM research trends. Section 6 considers potential future developments in this field. The paper concludes with Section 7 . 2 Background Traditionally, the combination of CMOS-based memories, such as volatile DRAM and SRAM, and non-volatile flash has been adequate to meet both the temporary and permanent data storage requirements of multi-chip systems. The trend towards system-on-chip integration with scalability, reconfigurability, and very low power has driven the development of additional eNVM technologies with new

Read the complete text