Dark Reading is part of the Informa Tech Division of Informa PLC

This site is operated by a business or businesses owned by Informa PLC and all copyright resides with them.Informa PLC's registered office is 5 Howick Place, London SW1P 1WG. Registered in England and Wales. Number 8860726.

Application Security

03:00 PM
Ari Singer
Ari Singer
Connect Directly
E-Mail vvv

TPM-Fail: What It Means & What to Do About It

Trusted Platform Modules are well-suited to a wide range of applications, but for the strongest security, architect them into "defense-in-depth" designs.

On November 12, researchers, led by a team at Worcester Polytechnic Institute, disclosed details of two new potentially serious security vulnerabilities — dubbed TPM-Fail — that could allow attackers to steal cryptographic keys protected inside two types of Trusted Platform Modules (TPMs): chips made by STMicroelectronics, and firmware-based Intel TPMs.

TPMs are deployed in billions of desktops, laptops, servers, smartphones, and Internet-of-Things (IoT) devices to protect encryption keys, identity keys, passwords, and system code. An attacker could use stolen cryptographic keys to forge digital signatures, steal or alter encrypted information, and bypass operating system security features or compromise applications that rely on the integrity of the keys.

Because millions of deployed systems probably have the TPM-Fail vulnerability, the scope of exposure is wide. It's especially troubling if companies and individuals don't update their firmware using patches now available from Intel and STMicroelectronics. Similar vulnerabilities may well exist in TPMs from other manufacturers, as well.

However, when this type of issue occurred in the past, the impact was largely contained. The fact that the patches were made available at the same time as the vulnerability was announced helps to minimize damage, particularly for security-conscious system owners. The challenge is that not everyone is ready to perform these patches whenever an exploit such as this becomes known.

This vulnerability demonstrates just how hard it is to make really strong security. Absolute security does not exist. Intel and ST are two very reputable companies that do things the right way: They get certifications, work with third parties, and promptly patch vulnerabilities when found. TPM-Fail reminds us that cybersecurity is always changing, and effective security against attacks today may not work against more sophisticated attacks tomorrow. Creating and implementing a very good security design is the start of a security process; the task of protecting devices and data is continually evolving.

Nation-States versus a 75-Cent Component
This can look like an unfair fight, tasking a component that is often priced under a dollar with protecting the most critical infrastructure and pitting it against the world's most sophisticated hacking experts. From politicians' smartphones to systems controlling military aircraft and servers holding your family's private data, much faith has been placed in TPMs.

Yet we have known for years that sophisticated attacks can be successful against devices such as these. Nobody should assume that a TPM or any other single security component can provide the "magic bullet" that protects a system against any and all compromises. We also need to recognize that some attacks, like this one, can leave the door open to less-sophisticated attackers taking advantage of vulnerabilities that someone else discovered to cause much more widespread damage.

Is somebody to blame? Probably not. A subtle point about these types of attacks, also called side-channel attacks, is that the implementations can be totally correct in terms of inputs and outputs, but information may still be leaked by observing the interaction of the device with the "real world." An example: measuring the time the device takes to perform specific crypto operations.

Commercial companies don't have unlimited budgets to make these components perfect, so they do their best with what they have, then take responsibility for addressing issues when they arise. In this case, STMicroelectronics and Intel designed their TPMs to be upgradeable and able to accept patches if an attack like this were to be found.

Encryption Key & TPM Safety
In most cases, encryption keys are safe. It's a fact, however, that any hardware device or firmware used for security purposes (e.g., TPMs and hardware security modules, or HSMs) may have design or firmware flaws that leave them susceptible. The attacks may expose critical cryptographic materials, including keys used to encrypt sensitive data. A recent example is the Roca vulnerability. Active monitoring of the latest vulnerabilities and patches should be used to make sure that your keys remain safe.

TPM-Fail should not cause everyone to stop using TPMs, but TPMs should not be seen as a "magical force field" for IoT and mobile devices. TPMs are one of the ways to provide roots of trust, and TPM-Fail does not mean their time is over. TPMs should be part of an overall security solution design and part of security in depth. In addition, because security requires constant vigilance and monitoring, your solution security design will need periodic scrutiny. The more cybersecurity experts you get to closely examine your security design, the better off you are.

Global Exposure to TPM-Fail Right Now?
Both Intel and STMicroelectronics provided prompt patches before the independent researchers broke the news. The patches should close those particular gaps and reduce (but not eliminate) the potential impact in the field. Practically speaking, however, many who own affected devices will be unaware of the TPM-Fail issue and it's likely that millions of devices will not be patched.

It is also worth noting how difficult and expensive it is to patch devices out in the field. Going forward, it would be wise for organizations to design anything that needs security with the ability to receive remote updates of device firmware — meaning secure updates where the remote device can only accept legitimate updates from a source it can verify as legitimate — and block everything else.

Think Cyber-Resilience and "Defense in Depth"
Most important, for the future, we need to design in the ability to recover when a device in the field is compromised. This capability is called cyber-resilience, defined as the ability of a CPU — and device — to continue functioning, or resume functioning, after an attack is detected. Take driverless cars, for example. It's not acceptable if a component in the steering system is attacked and goes offline while the car is flying down a highway.

TPM-Fail also makes a strong argument for "defense in depth" because even the better commercial offerings can be broken. So, put layers of protection in place, including new chip and firmware offerings that can add defense in depth to single-depth security solutions. That way, attackers must find their way through multiple security mechanisms.

A final recommendation: Have a process in place to monitor and react to newly announced security vulnerabilities, including the ability to securely field-upgrade the firmware on devices, and install patches without delay.

Check out The Edge, Dark Reading's new section for features, threat data, and in-depth perspectives. Today's top story: "How Medical Device Vendors Hold Healthcare Security for RansomSoft Skills."

Ari Singer, CTO at TrustiPhi and long-time security architect with over 20 years in the trusted computing space, is former chair of the IEEE P1363 working group and acted as security editor/author of IEEE 802.15.3, IEEE 802.15.4, and EESS #1. He chaired the Trusted Computing ... View Full Bio

Recommended Reading:

Comment  | 
Print  | 
More Insights
Threaded  |  Newest First  |  Oldest First
User Rank: Apprentice
11/19/2019 | 10:06:35 PM
thank so much
Thank Pro
COVID-19: Latest Security News & Commentary
Dark Reading Staff 9/25/2020
Hacking Yourself: Marie Moe and Pacemaker Security
Gary McGraw Ph.D., Co-founder Berryville Institute of Machine Learning,  9/21/2020
Startup Aims to Map and Track All the IT and Security Things
Kelly Jackson Higgins, Executive Editor at Dark Reading,  9/22/2020
Register for Dark Reading Newsletters
White Papers
Current Issue
Special Report: Computing's New Normal
This special report examines how IT security organizations have adapted to the "new normal" of computing and what the long-term effects will be. Read it and get a unique set of perspectives on issues ranging from new threats & vulnerabilities as a result of remote working to how enterprise security strategy will be affected long term.
Flash Poll
How IT Security Organizations are Attacking the Cybersecurity Problem
How IT Security Organizations are Attacking the Cybersecurity Problem
The COVID-19 pandemic turned the world -- and enterprise computing -- on end. Here's a look at how cybersecurity teams are retrenching their defense strategies, rebuilding their teams, and selecting new technologies to stop the oncoming rise of online attacks.
Twitter Feed
Dark Reading - Bug Report
Bug Report
Enterprise Vulnerabilities
From DHS/US-CERT's National Vulnerability Database
PUBLISHED: 2020-09-25
In tensorflow-lite before versions 1.15.4, 2.0.3, 2.1.2, 2.2.1 and 2.3.1, when determining the common dimension size of two tensors, TFLite uses a `DCHECK` which is no-op outside of debug compilation modes. Since the function always returns the dimension of the first tensor, malicious attackers can ...
PUBLISHED: 2020-09-25
In tensorflow-lite before versions 1.15.4, 2.0.3, 2.1.2, 2.2.1 and 2.3.1, a crafted TFLite model can force a node to have as input a tensor backed by a `nullptr` buffer. This can be achieved by changing a buffer index in the flatbuffer serialization to convert a read-only tensor to a read-write one....
PUBLISHED: 2020-09-25
In tensorflow-lite before versions 1.15.4, 2.0.3, 2.1.2, 2.2.1 and 2.3.1, if a TFLite saved model uses the same tensor as both input and output of an operator, then, depending on the operator, we can observe a segmentation fault or just memory corruption. We have patched the issue in d58c96946b and ...
PUBLISHED: 2020-09-25
In TensorFlow Lite before versions 1.15.4, 2.0.3, 2.1.2, 2.2.1 and 2.3.1, saved models in the flatbuffer format use a double indexing scheme: a model has a set of subgraphs, each subgraph has a set of operators and each operator has a set of input/output tensors. The flatbuffer format uses indices f...
PUBLISHED: 2020-09-25
In TensorFlow Lite before versions 2.2.1 and 2.3.1, models using segment sum can trigger writes outside of bounds of heap allocated buffers by inserting negative elements in the segment ids tensor. Users having access to `segment_ids_data` can alter `output_index` and then write to outside of `outpu...