Hardware backdoor

Lore & Background

The existence of hardware backdoors poses significant security risks for several reasons. They are difficult to detect and are impossible to remove using conventional methods like antivirus software. They can also bypass other security measures, such as disk encryption. Hardware trojans can be introduced during manufacturing where the end-user lacks control over the production chain.

In 2008, the FBI reported the discovery of approximately 3,500 counterfeit Cisco network components in the United States, some of which were introduced in military and government infrastructure. In the same year, the possibility of a backdoor SPARC CPU was demonstrated with an FPGA running Linux that supported various hidden malicious services. A few years later, in 2011, Jonathan Brossard presented 'Rakshasa', a proof-of-concept hardware backdoor that could be installed by an individual with physical access to the hardware. It utilized coreboot to re-flash the BIOS with a SeaBIOS and iPXE-based bootkit composed of legitimate, open-source tools, allowing malware to be fetched from the internet during the boot process.

In 2012, Sergei Skorobogatov and Christopher Woods from the University of Cambridge Computer Laboratory reported the discovery of a backdoor in a military-grade FPGA device, which could be exploited to access and modify sensitive information. It has been said that this was proven to be a software problem and not a deliberate attempt at sabotage. Later that year, two mobile phones developed by the Chinese company ZTE were found to carry a root access backdoor using a hard-coded password. Starting in 2012, the United States stated that Huawei might have backdoors present in their products. In 2013, researchers at the University of Massachusetts devised a method of breaking a CPU's internal cryptographic mechanisms by introducing specific impurities into the crystalline structure of transistors to change Intel's random-number generator. Documents revealed from 2013 onwards during the surveillance disclosures initiated by Edward Snowden showed that the Tailored Access Operations (TAO) unit and other NSA employees intercepted servers, routers, and other network gear being shipped to organizations targeted for surveillance to install covert implant firmware onto them before delivery.

Reader's Guide

The significance of hardware backdoors lies in their ability to undermine security at a fundamental level, as they are difficult to detect and impossible to remove using conventional methods like antivirus software. They can bypass other security measures, such as disk encryption. The article highlights that backdoors are introduced intentionally by the original designer or during the design process, distinguishing them from hardware Trojans inserted later by an external party. This intentional nature makes them particularly insidious, as they can be embedded in critical infrastructure, as seen with the 2008 discovery of counterfeit Cisco components in military and government systems. The legacy of hardware backdoors includes ongoing efforts to detect and mitigate them. Countermeasures described in the article include Skorobogatov's technique for detecting malicious insertions into chips, verifiable computing methods from New York University Tandon School of Engineering that embed verification modules in chips, and a technique from University College London that distributes trust between multiple identical chips from disjoint supply chains. Additionally, researchers at the University of Southern California and the Paul Scherrer Institute developed Ptychographic X-ray laminography, a non-destructive method for verifying a chip's blueprint and design, which can identify the foundry and aspects of the design. These countermeasures reflect the ongoing arms race between those who create hardware backdoors and those who seek to detect and prevent them.

Did You Know?

• Hardware backdoors are intentionally introduced by the original designer or during the design process, unlike hardware Trojans which are inserted later by an external party.
• In 2016, University of Michigan researchers built a hardware backdoor using analog circuits that could be activated after capacitors charged, giving complete access to the system.
• The 2011 Rakshasa proof-of-concept backdoor used coreboot to re-flash the BIOS with a bootkit that fetched malware from the internet during boot.
• Ptychographic X-ray laminography is the only current method for non-destructive reverse engineering of chips to verify they are manufactured according to design.

Definition and Nature

A hardware backdoor is a vulnerability deliberately embedded within the physical components of a computing system, residing in the hardware layer rather than in software. These can be introduced by writing malicious code into a component's firmware or even during the fabrication of an integrated circuit. Their most common targets are smartcards and cryptoprocessors, though researchers have also explored their application in automotive contexts. A key distinction separates backdoors from hardware Trojans: backdoors are placed intentionally by the original designer or during the design phase, whereas Trojans are inserted later by an external actor. This intentional integration makes them especially dangerous, as they become architecturally inseparable from the device. Conventional defenses such as antivirus programs are entirely ineffective against them, and they can bypass protections like disk encryption. Because end-users have no oversight over the manufacturing chain, a backdoor planted at the factory level remains invisible and cannot be removed through standard means.

Notable Incidents and State-Level Exploitation

The history of hardware backdoors is punctuated by alarming real-world discoveries. In 2008, the FBI uncovered roughly 3,500 counterfeit Cisco network components in the United States, some of which had been deployed in military and government infrastructure. That same year, researchers demonstrated a backdoored SPARC CPU using an FPGA running Linux that hosted hidden malicious services. In 2011, Jonathan Brossard presented Rakshasa, a proof-of-concept backdoor installable by anyone with physical access, which used coreboot to re-flash a BIOS with a SeaBIOS and iPXE-based bootkit built from legitimate open-source tools, enabling malware retrieval from the internet at boot. In 2012, Cambridge researchers found a backdoor in a military-grade FPGA, and two ZTE mobile phones were discovered carrying a root-access backdoor with a hard-coded password. From 2012 onward, the United States raised concerns about potential backdoors in Huawei products. Snowden's 2013 disclosures revealed that the NSA's Tailored Access Operations unit intercepted servers, routers, and network gear in transit to implant covert firmware, including custom BIOS exploits and USB cables packed with spy hardware and radio transceivers.

Research Frontiers and Proof-of-Concept Attacks

Academic research has repeatedly demonstrated how feasible and stealthy hardware backdoors can be. In 2013, University of Massachusetts researchers devised a method of subverting a CPU's internal cryptographic mechanisms by introducing specific impurities into the crystalline structure of transistors, thereby altering Intel's random-number generator. In June 2016, the University of Michigan's Department of Electrical Engineering and Computer Science built a backdoor leveraging analog circuits: once capacitors within the chip accumulated enough charge, the circuit would activate, granting an attacker full access to the host system. Their study, which won the best paper award at the IEEE Symposium on Privacy and Security, noted that such a microscopic implant would evade virtually every modern hardware security analysis method and could be planted by a single factory employee. In October 2018, Bloomberg reported that Chinese state-sponsored actors had compromised the U.S. technology supply chain to reach nearly 30 American companies, including Amazon and Apple, illustrating that these academic concerns had already translated into operational threats at scale.

Countermeasures and Detection Techniques

Defending against hardware backdoors requires specialized approaches, since standard software security tools are useless. Sergei Skorobogatov developed a technique capable of detecting malicious insertions directly within chips. At New York University's Tandon School of Engineering, researchers created a verifiable computing framework in which chips sold for commercial use carry an embedded verification module that attests to the correctness of their calculations, validated by an associated external module. University College London researchers proposed distributing trust across multiple identical chips sourced from disjoint supply chains, so that as long as at least one chip remains honest, device security is preserved. Perhaps most promising is Ptychographic X-ray laminography, developed jointly by the University of Southern California and the Paul Scherrer Institute. It is the only current method that verifies a chip's blueprint without destroying or cutting it, doing so far faster than alternatives. Professor Anthony F. J. Levi described it as a fingerprint for non-destructive reverse engineering. It can scan chips up to 12 by 12 millimeters in 3D, easily accommodating an Apple A12 chip, though not yet a full Nvidia Volta GPU.