Embedded Systems Security

I’ve discovered a significant vulnerability in the Arm® TrustZone® CryptoCell 310 AES-128 hardware engine within the Nordic Semiconductor nRF52840 SoC, which has been acknowledged in the Nordic Security Advisory SA-2025-380-v1.0. By inducing precisely timed voltage fault injection, I was able to bypass the AES encryption and recover plaintext in ECB, CBC, and CTR modes. This vulnerability exposes sensitive data, making cryptographic methods like key rotation and CBC initialization vectors ineffective. Additionally, observed error byte diffusion under fault conditions could be potentially exploited for Differential Fault Analysis (DFA), particularly in ECB mode. This is a hardware vulnerability that cannot be patched easily without a redesign of the silicon. However, I proposed several firmware-level countermeasures to mitigate the attack's impact. These techniques, while effective, come with performance trade-offs and should be evaluated based on the specific use case. The attack was performed using my custom low-cost voltage fault injection tool based on the crowbar technique. As physical access is required for such attacks, minor hardware modifications like capacitor removal are not a significant barrier. Still, I achieved successful results without needing to remove any capacitors, demonstrating the practicality and repeatability of this approach. Discovery Reported to Nordic Semiconductor ASA on Nov 13, 2024. Public Disclosure Coordinated with Vendor on Apr 11, 2025. Full Research: https://lnkd.in/dZuMmp2C Nordic Security Advisory: SA-2025-380-v1.0 https://lnkd.in/dNCrabzK Sharing this to inform and support the embedded and hardware security community. #cybersecurity #infosecurity #hardwaresecurity #penetrationtesting #cyberdefense #cyberattack #redteam #redteaming #vulnerabilities #cryptography #hardwaredesign #hardwareengineering #softwareengineering #iotdevices #iotsecurity #hardware #development #communication #embeddedsystems #embeddedsoftware #iot

Analysts across Gartner, Forrester, and Frost all highlight the same frontier: Unmanaged, un-agentable, unpatchable legacy and embedded mission-critical devices. These systems now power hospitals, factories, transportation networks, logistics hubs, energy grids, and smart buildings. And they break every assumption that traditional cybersecurity was built on. You can’t install an agent, can’t run a scanner, can’t take downtime to patch, can’t modify the configuration without operational impact, and you often can’t replace the device, even when it’s vulnerable. This isn’t an edge case anymore. It’s the dominant surface. Traditional IT security assumes: → You own the device → You can instrument it → You can patch it → You can enforce controls → You can model its behavior None of this holds in cyber-physical systems. Industrial controllers run firmware older than some of the engineers maintaining them. ↳ Medical devices can’t be scanned because they may disrupt patient care. ↳ Building automation systems weren’t designed with authentication in mind. ↳ Robotics and sensors can’t tolerate downtime. ↳ IoT devices run proprietary protocols no EDR understands. And yet these devices are connected: to your network, your cloud, your authentication systems, and your business processes. This is where attackers are moving their focus. Modern CPS protection platforms increasingly rely on three capabilities analysts repeatedly emphasize: 1. Passive, protocol-level discovery Visibility without disruption. Understanding devices based on their behavior. 2. Contextual exposure analysis → What is reachable? → What is in the attack path? → What would cause operational impact? This is the only way to prioritize unpatchable devices. 3. Compensating controls instead of patching When patching is impossible, risk reduction happens through: Segmentation, policy enforcement, traffic shaping, identity hardening, control-plane exceptions, behavioral monitoring, and attack-path suppression This is where modern architectures are now moving. Across industries, leaders are accepting a new reality: We won’t regain control of these devices. We must build control around them. That’s why CPS protection platforms now combine: Asset intelligence, network-centric defenses, reachability mapping, risk scoring, MITRE ICS/TTP alignment, operational workflow, integrations with IT security, and AI-based anomaly detection. It’s becoming the backbone of securing the parts of the enterprise you can’t instrument. The new frontier of security is securing environments built from devices you never truly controlled in the first place.

ESP32-based IoT devices are often deployed in physically accessible, network-connected, and long-lived environments. These characteristics make them attractive targets for attackers seeking persistent access, device cloning, or data exfiltration. Unlike traditional IT systems, embedded devices cannot rely on perimeter defenses alone. Security must be built into the firmware, boot process, and hardware configuration from the first instruction executed. This article presents practical, field-tested defense strategies for securing ESP32-based embedded systems using ESP-IDF. The focus is not on abstract security theory, but on concrete mechanisms available in real ESP32 silicon: secure boot, flash encryption, eFuses, TLS, and secure OTA workflows. Each section explains what problem the mechanism solves, why it matters, and how to implement it correctly. #learningbytutorials #esp32 #esp32projects #espidf #embeddedsystems #embeddedprogramming

A flaw in Infineon’s security microcontrollers made it possible to extract secret keys using a lab setup that cost just $11,000. 📟🔑👊🏻👨💻 A few months ago, security researcher Thomas Roche presented his fundamental research on secure elements used in the YubiKey 5. The security element is the Infineon SLE78, which contains a proprietary implementation of the Elliptic Curve Digital Signature Algorithm (ECDSA). Using side-channel attacks and a great deal of smart research, the author discovered a vulnerability in Infineon Technologies' cryptographic library and, as a result, was able to extract the ECDSA secret key from the secure element. The cost of the setup was €10,000, including the laptop. Let me quote the author: "...in fact, all Infineon security microcontrollers (including TPMs) that run the Infineon cryptographic library (as far as we know, any existing version) are vulnerable to the attack." Infineon is one of the most popular manufacturers of secure elements across many industries, including: 🔮 Automotive - used for SecOC and V2X key storage 🔮 Medical - used for secure communication, device pairing, and patient data storage 🔮 OT (Operational Technology) - used to ensure secure data transmission and device authentication 🔮 Avionics - used to ensure firmware integrity, protect IFEC systems, and enable secure communication with ground systems ...and more. Please stay safe and share this with your peers responsible for security and safety. It's important for them to be informed. More details: Side-Channel Attack on the YubiKey 5 Series [PDF]: https://lnkd.in/dvPjUV4R #hacking #embedded #Infineon #ECDSA #TPM #security #safety #cyber #tech #technology #YubiKey #privacy #attack #medical #automotive #avionics #SCADA #IoT

The ability to update software on devices is a valuable tool for protecting critical systems from evolving threats. However, this capability is not without risk. There have been an alarming number of vulnerabilities that were introduced through a malicious software patch or a flaw in the update process. New software update frameworks have been developed to mitigate this risk, but they come with new levels of complexity, and they may not work on segmented network architectures or be suitable for embedded devices. Brian Romansky focuses on TUF (The Update Framework), a software update approach that addresses many common vulnerabilities and consider how it can be applied in a critical infrastructure environment. It is compared against SUIT (Software Update for IoT) and UpKit, two alternative structures that are intended for use on embedded systems. Attack trees are used to compare these models and visually explain the strengths and challenges that may be encountered when they are applied in a network that follows the Purdue or ISA-99/IEC 62443 network architecture. The role of metadata such as an SBOM and vendor test results are also be considered. These concepts are merged to re-cast software updates into the context of an integrated supply-chain and configuration management system.

Modern rolling stock carries hundreds of sensors, embedded controllers, and connected systems that interact with signaling, passenger Wi-Fi, ticketing, and maintenance networks. This evolution has improved efficiency and passenger comfort, but it has also opened a new cyber battleground. Attacks that were once aimed at back-office IT systems now target train control systems, onboard diagnostics, and even communication protocols like GSM-R and its successor, FRMCS. The railway sector has already seen wake-up calls. In 2022, a ransomware attack on a regional train operator forced service delays and manual traffic control. In 2024, a vulnerability disclosure showed that insecure firmware updates on onboard controllers could allow remote manipulation of braking systems. These incidents illustrate that railway cybersecurity is no longer hypothetical; it is a real operational risk. Resilience starts with architecture. Segmenting train networks is critical, separating passenger Wi-Fi and infotainment systems from safety-critical control domains, and isolating signaling communication from external entry points. The IEC 62443 framework provides a strong foundation, defining zones and conduits that restrict access and limit lateral movement. EN 50159 and TS 50701 add railway-specific guidance, covering secure transmission protocols and lifecycle security management tailored to signaling and rolling stock. Zero Trust principles are increasingly being applied to railway operations, verifying identities and device health before granting access to critical systems. Strong encryption, secure boot, and signed firmware updates are essential to protect embedded devices from tampering. Additionally, the use of intrusion detection tailored to operational technology networks is helping operators detect malicious activity quickly, even in environments where patching cycles are slower due to safety certification constraints. Another critical layer is supply chain assurance. Rolling stock manufacturers depend on a complex network of component suppliers, and a compromised subsystem can introduce vulnerabilities that bypass perimeter defenses. Security audits, SBOMs (Software Bill of Materials), and contractual security requirements are becoming standard to manage this risk. Looking forward, the integration of FRMCS, the next-generation mobile communication system for rail, adds both opportunity and complexity. While FRMCS offers stronger encryption and flexible bandwidth, its IP-based architecture increases exposure to internet-style attacks. Proactive measures, like continuous monitoring, red teaming, and vulnerability disclosure programs, will be key to staying ahead. Railway operators, infrastructure managers, and manufacturers must treat cybersecurity as part of operational safety. The line between digital and physical security has blurred. #RailwaySecurity #CyberResilience #RollingStock #OTSecurity #IEC62443 #EN50159 #TS50701 #CriticalInfrastructure

What if I told you that C++26 could eliminate every malloc() and free() call from your embedded IoT stack while making it MORE secure? Summary: Just published my latest deep-dive into how C++26's static reflection can revolutionize embedded protocol development. Using CoAP (Constrained Application Protocol) as a real-world case study, I demonstrate: ✅ Zero dynamic allocation - Complete protocol implementation with compile-time memory determination ✅ Compile-time security policies - Field-level encryption annotations that prevent data leaks by design ✅ DTLS integration - Secure IoT communication without runtime overhead ✅ Safety-critical compliance - Ready for DO-178C, ISO 26262, and IEC 62443 certification The article shows how a temperature sensor CoAP server can be built with deterministic behavior, automatic serialization, and provable security - all generated at compile time. This isn't just academic theory - it's production-ready techniques for the next generation of connected embedded devices where safety and security can't be an afterthought. Call-to-Action: Full technical implementation details, code examples, and security patterns in the article linked below. What embedded challenges are you tackling that could benefit from compile-time guarantees? Hashtags: #CPP26 #EmbeddedSystems #StaticReflection #CoAP #EmbeddedCPlusPlus #SafetyCritical #IoT #DTLS #CompileTime #ZeroOverhead #EmbeddedSecurity #ModernCPlusPlus #EmbeddedDevelopment

The main principle I teach about DevSecOps in embedded systems is this: Shift Security Left. If you imagine your development cycle as a left-to-right timeline, most teams push security all the way to the right. They treat it like a final checkbox, not a design constraint. Shifting security left means doing the opposite: deliberately pulling security practices into the early stages of development. You might think, "Whoa, doesn’t that mean shipping a product will take longer and cost more?" Good question. And the answer surprisingly is a big fat NO. Yes, integrating security early forces more testing. Yes, it can feel repetitive and tedious. Yes, it adds friction where most teams prefer comfort. But that’s exactly the point: That friction is so painful if done manually, it forces developers to automate. And automation will lead to less time and less money wasted. That’s why the second principle I teach after “Shift Security Left” is to learn and adopt automation. You see, those early checks surface vulnerabilities when they’re still cheap to fix. And that’s how teams who shift security left end up moving faster, not slower.

I’ve been having a lot of fun working on the material for my upcoming Embedded Security training - and I’m happy to say it’s almost 70% done. 😀 The training is divided into 4 parts and 12 topics, to be delivered over four days (3 topics per day). In the first day, we will start with a quick introduction to security fundamentals and threat modeling, and then dive into two big secure coding topics. In the first topic, I plan to cover secure coding from the attacker's perspective. The plan is to answer the question: how do vulnerabilities actually get found and exploited? I'll cover a wide range of techniques and tools: static analysis, sanitizers, fuzzing, Valgrind, reversing, GDB, binwalk, Ghidra, and more. We’ll even write a small shellcode to exploit a buffer overflow (strictly for educational purposes, of course!). In the second topic, I plan to cover secure coding from the defender’s perspective. The plan is to answer the question: how do we avoid vulnerabilities and harden systems against exploitation? Here we look at how Rust helps build safer code, security-focused coding standards, static analysis (again), Fortify, stack protector, RELRO, ASLR, pointer authentication, kernel hardening, and many others! And this is just on the first day! If you’re curious, the full agenda is here: https://lnkd.in/dAHnJh5D And if you can guess what I'm running in each of the five windows in the screenshot, drop a comment! :-) #security #embeddedsecurity #staticanalysis #sanitizers #fuzzing #reversing #rust #hardening