Victims of defective products are not required to demonstrate the “fault” of a product manufacturer. It’s enough to demonstrate the existence of a defect in the product that causes harm. Under European laws, “a product is defective when it does not provide the safety which a person is entitled to expect taking all circumstances into account…”  at Art.6;  at s.3.
Product strict liability has always been a source of concern for manufacturers (and importers, who are subject to the same liability). They are obviously concerned about liability in the absence of fault. Unlike many other forms of liability (like warranty), manufacturers are practically unable to limit this liability to victims who sue alone or collectively in a class action.
Two important conditions must exist before a victim can succeed on a strict liability claim: 1) There must be a “product” which is defective
2) A victim harmed by a defective product can only use this legal theory to claim compensation for death or personal injury (or damage to non-commercial property under the laws of the EU). Economic harm, business interruption, loss of business revenue, etc, are not recoverable under this theory. These two conditions made strict liability a niche topic or an intellectual curiosity for most lawyers working in the fields of software development and cyber security and meant that it was traditionally overlooked in these fields. For decades we have taken comfort in the widely shared legal opinion that software, as such, does not fit within the definition of “product” under European or American laws. Even if software was to be viewed as a product, we reasoned, opportunities for defective software design to cause death or personal injury seemed exceedingly rare. One long-understood risk of strict liability concerns defective software control systems as a component in safety-critical hardware. The manufacturer of the resulting defective hardware is subject to strict liability claims, irrespective of the source of the defect. This risk can be illustrated with the example of the Therac-25 radiation therapy machine. Between 1985-87, six patients treated using the Therac-25 were exposed to massive radiation overdoses (100x intended dose). Three of these patients died as a result of the overdoses. The design of the machine’s system control software is widely cited as a cause of the overdose incidents, which were thankfully rare.  Under a strict liability analysis, the Therac-25 device as a whole is a “product”. If the machine failed to provide the “safety which a person is entitled to expect,” such a product would be defective and the manufacturer strictly liable for personal injury or death. The fact that the flaw originated in control software would be irrelevant. For decades, my legal colleagues and I rested comfortable in the belief that software errors (including software security flaws) rarely killed anyone. Today, by contrast, the IoT presents a rapidly growing set of opportunities for “death by software”. A net-connected software-controlled product (e.g., an autonomous vehicle, an industrial control system, a pacemaker, a vehicle using fly-by-wire) that fails to deliver appropriate safety, is defective whether the safety is compromised through the design of electrical, mechanical, software, or security, systems. Thus strict liability applies to products whether safety is compromised through errors in algorithmic decision-making (e.g., an autonomous vehicle decides to swerve into oncoming traffic after misreading road markings) or security errors (e.g., a broken authentication scheme permits a remote hacker to divert the same vehicle into oncoming traffic). While the hardware product manufacturer (or importer) is clearly subject to the risk of strict liability, what about those in the upstream supply chain? What if, for example, the manufacturer of the Therac-25 had purchased their control software from a third party as a component, or the autonomous vehicle manufacturer adopts and installs a defective authentication package embodied in third-party software? Under current law, defective component “product” manufacturers face strict liability. A manufacturer of defective brakes, for example, is strictly liable for personal injury caused by automobiles which become defective because the defective brakes are installed. Software (on its own) is not currently thought to be a product in this area of law. The author of a defective software component probably cannot face a strict liability claim from an injured victim – even if the software caused the hardware product to harm the victim. This may be about to change. More than three decades have passed since the 1985 adoption of the European Directive on product strict liability . The reliance society places on software and online services has become a central feature of everyday life. European policy makers have noticed, and the tide of product liability policy appears to be shifting. The European Commission completed a comprehensive evaluation of European product liability law in 2018. The term “software” features prominently, and repeatedly, in the 108-page report . The Commission openly questions the extent to which “digital products” (e.g., software as a product, SaaS, PaaS, IaaS, data services, etc.) should be redefined as “products” and thus subjected to strict liability analysis when defects cause death or personal injury . A Commission Expert Group on liability and new technologies is currently examining possible changes to the law. Expanding the definition of “product” is central to this review. We seem to be accelerating towards a world in which cyber security failures in the IoT will create increasing risk to life and limb. Manufacturers of tangible IoT products already face strict liability if their product is unsafe – including cases where safety is compromised by poor cyber security. It appears that software developers, SaaS providers, and other cloud service providers, may soon be required to step up to this same stringent standard of responsibility throughout Europe. We hope they’ll be prepared for the challenge. Works Cited:  European Economic Community, Council Directive of 25 July 1985 on the approximation of the laws, regulations and administrative provisions of the Member States concerning liability for defective products (85/374/EEC), vol. L210, 1985, p. 29.
 Consumer Protection Act 1987.
 N. Leveson, “Medical Devices: The Therac-25,” in Safeware: System Safety and Computers, Addison-Wesley, 1995.
 European Commission, Evaluation of Council Directive 85/374/EEC of 25 July 1985 on the approximation of the liability for defective products, Brussels, 2018.
 European Commission, Liability for emerging digital technologies, Brussels, 2018.
In this article, we briefly outline the challenges faced by regulated organisations that need procedural control over their HSM estate. It provides an insight into the level of awareness, lack of resources and expertise that pose significant challenges in production system adoption.
The need for procedures
Regulation dominates the payment systems behind global financial institutions. Primarily because most financial service providers are regulated to a standard by typically a local payments network provider such as LINK UK or an international provider such as VISA or Mastercard. In both cases regulatory control such as the Payment Card Industry (PCI) PIN is applicable either because PCI PIN is mandated directly by the international providers regulators (the PCI council) or because the local network provider adopts, legislates and regulates the relevant controls mandated in the PCI standard.
Regulatory compliance is at the forefront of all payment service provider’s service enablement strategy. It is the primary reason why the technical and procedural controls are in place. Failure to comply with the regulatory requirements through poor procedures could lead to steep penalties levied by the card brands and suspension of network use or card issuance for the issuer. Such a threat means that card issuers and switch service providers are highly sensitive to any aspect of the regulation not being adhered to and poor procedures are a factor. Furthermore poor procedures could lead to a costly security breach followed by the risk of reputational damage.
A significant aspect of regulatory control is the development and demonstrable use of procedures. PCI PIN Security mandates that virtually every control objective related to keys, components or HSMs has a policy and procedure. Furthermore, the business top down ownership of the procedures must be in place and all affected parties (key custodians, supervisory staff, technical management, etc.) must be aware of those procedures. Without procedures compliance is simply not possible and the fact is that most business owners don’t really know where they stand with the state of their procedures until it is too late, furthermore procedure custodians feel vulnerable with their liabilities at audit time.
HSM related procedures enable control and consistency over the way a company achieves an objective as well as to demonstrate retrospective control over its device, component and key management activities sometimes long after the activity has taken place. Crucially, signed procedures are the only way to attest for an event post performance to an auditor and so attested performances must be meticulously planned and coordinated to correctly capture any event. 61% of respondents in the 2019 Ponemon Global Encryption Trends Survey states that key management is painful with the top reasons shown below for which the right procedures can help:
The industry faces a common problem in finding it difficult to create and maintain a reasonably reliable quality programme for the creation of procedures, this issue needs to be addressed.
Personal experience has witnessed the fact that procedures often do not get enough visibility until very late on in a project and that little time, budget and resource allocation is assigned to them. The lack of resources given to the creation of procedures means that they are typically:
- Not demonstrative because they don’t sufficiently detail the process
- Do not capture the right sign off at the relevant points and therefore cannot attest to the performance
This lack of detail and attestation leaves little for an auditor to validate and without procedures there is a strong chance that the platform simply won’t make production.
The irony is that businesses all over the world, using the same HSM products to satisfy the same regulation for interaction with similar card schemes require carbon copies of procedures but without help they face varying problems. This help is not available because vendors are interested in product delivery and businesses are interested in service delivery, very little support is available to solve the problems that exist in the gaps.
Procedure creation is typically left to custodians within the business with little understanding of the intricacies of HSM implementation. In some cases an HSM was a reluctantly inherited item of hardware that in some cases was considered as simply an encryption router and treated as such.
Having worked for a vendor for many years as an HSM SME, my consulting role was typically to focus on the implementation of the HSM into its environment and any associated “near box” activities. I could be involved with pre-production or test, but it was unlikely that I would ever be part of the wider needs of standards such as ITIL service delivery. Vendor consultants typically do not assist with high or low lever designs, implementation planning, creation of logs for component access, device inventories, physical inspections or procedure development.
Consideration of the wider service delivery requirement provides a better experience in the handover of an HSM to the business from project to support, anything less is a problem which needs addressing.
Procedure development requires expertise. Having certified on a 2-day HSM capabilities training course and passing the exam is a great start to gaining that expertise but procedure-based training on how to complete daily operations is better. Access to a costly spare HSM is also fantastic to refresh skills and to get familiar with various controls but access to the required wisdom to make informed decisions on procedure development or to have them developed for you is better.
To summarise, some reasons for poor procedures in an organisation are listed:
- They are not backed by a Security Policy or sufficiently business owned
- They are not given enough time or budget for success
- Lack of hands-on HSM experience means that they are created without correct workflow and therefore not accurate
- They don’t reflect new product updates
- They can be inconsistent and disjointed due to multiple authors and no peer review
- They are created without a reasonable knowledge of the regulatory requirements
- They are created without any consideration of the vendors security recommendations
- The organisation is unclear of how to change a procedure after an auditor’s non-compliance notice
- Issues of versioning mean that the procedure is based on legacy features
- The procedures typically contain missing or incomplete signatories and attestation
- Paper based procedures become disorganised, lost, incomplete as well as hard to locate when requested
- On multiple occasions the procedures are not available for the performance
All of the above make HSM procedure management problematic for businesses globally. The business is accountable for the regulation and it will be the business that receives the non-compliance when a procedure does not satisfy its regulatory requirements. Put simply, HSM procedure mismanagement is a ticking time bomb for the unaware and should be considered a threat that could give rise to significant risk that must be managed. Not to take control over procedure management is naive at best and negligent at worst. Organisations need to regularly review how they are managing the risk or poor procedures.
With the increasing dependence on cryptography to protect digital assets and communications, the ever-present vulnerabilities in modern computing systems, and the growing sophistication of cyber attacks, it has never been more important, nor more challenging, to keep your cryptographic keys safe and secure. A single compromised key could lead to a massive data breach with the consequential reputational damage, punitive regulatory fines and loss of investor and customer confidence.
In this article we look at why cryptographic keys are one of your company’s most precious assets, how these keys can be comprised, and what you can do to better protect them – thereby reducing corporate risk and enhancing your company’s cyber-security posture.
Cryptography lies at the heart of the modern business – protecting electronic communications and financial transactions, maintaining the privacy of sensitive data and enabling secure authentication and authorization. New regulations like GDPR and PSD2, the commercial pressure for digital transformation, the adoption of cloud technology and the latest trends in IoT and blockchain/DLT all help drive the need to embed cryptography into virtually every application – from toasters to core banking systems!
The good news is that modern cryptographic algorithms, when implemented correctly, are highly-resistant to attack – their only weak point is their keys. However, if a key is compromised, then it’s game over! This makes such cryptographic keys one of your company’s most precious assets, and they should be treated as such. The value of any key is equivalent to the value of all the data and/or assets it is used to protect.
There are three primary types of keys that need to be kept safe and secure:
- Symmetric keys – typically used to encrypt bulk data with symmetric algorithms like 3DES or AES; anyone with the secret key can decrypt the data
- Private keys – the secret half of public/private key pairs used in public-key cryptography with asymmetric algorithms like RSA or ECDSA; anyone with the private key can impersonate the owner of the private key to decrypt private data, gain unauthorized access to systems or generate a fraudulent digital signature that appears authentic
- Hash keys – used to safeguard the integrity and authenticity of data and transactions with algorithms like HMAC-SHA256; anyone with the secret key can impersonate the originator of the data/transactions and thus modify the original data/transactions or create entirely false data/transactions that any recipient will believe is authentic
With an ever-increasing number of keys to protect, and an ever-increasing value of data being protected by those keys, not to mention the demands of PCI-DSS or GDPR, this is a challenge that nearly every business needs to face and address as a matter of urgency.
What dangers await?
There are many threats that can result in a key being compromised – typically, you won’t even know the key has been compromised until it has been exploited by the attacker, which makes the threats all the more dangerous. Here are some of the major threats that should be considered:
A key is essentially just a random number – the longer and more random it is, the more difficult it is to crack. The strength of the key should be appropriate for the value of the data it is protecting and the period of time for which it needs to be protected. The key should be long enough for its intended purpose and generated using a high-quality (ideally certified) random number generator (RNG), ideally collecting entropy from a suitable hardware noise source.
There are many instances where poor RNG implementation has resulted in key vulnerabilities.
Incorrect use of keys
Each key should be generated for a single, specific purpose (i.e. the intended application and algorithm) – if it is used for something else, it may not provide the expected or required level of protection.
Re-use of keys
Improper re-use of keys in certain circumstances can make it easier for an attacker to crack the key.
Non-rotation of keys
If a key is over-used (e.g. used to encrypt too much data), then it makes the key more vulnerable to cracking, especially when using older symmetric algorithms; it also means that a high volume of data could be exposed in the event of key compromise. To avoid this, keys should be rotated (i.e. updated / renewed) at appropriate intervals.
Inappropriate storage of keys
Keys should never be stored alongside the data that they protect (e.g. on a server, database, etc.), as any exfiltration of the protected data is likely to compromise the key also.
Inadequate protection of keys
Even keys stored only in server memory could be vulnerable to compromise. Where the value of the data demands it, keys should be encrypted whenever stored and only be made available in unencrypted form within a secure, tamper-protected environment and even (in extreme cases) kept offline.
Insecure movement of keys
It is often necessary to move a key between systems. This should be accomplished by encrypting (“wrapping”) the key under a pre-shared transport key (a key encryption key, or KEK), which may be either symmetric or asymmetric. Where this is not possible (e.g. when sharing symmetric transport keys to bootstrap the system), the key should be split into multiple components that must then be kept separate until being re-entered into the target system (and then the components are destroyed).
Non-destruction of keys
Keys should be destroyed (i.e. securely deleted, leaving no trace) once they have expired, unless explicitly required for later use (e.g. to decrypt data). This removes the risk of accidental compromise at some future date.
Insider threats (user authentication, dual control, segregation of roles)
One of the biggest classes of threat that a key faces is insider threats. If a rogue employee has unfettered access to a key, they might use it for a malicious purpose or pass it onto someone else to the same end.
Lack of resilience
Not only must the confidentiality and integrity of keys be protected, but also their availability. If a key is not available when required, or worse still lost due to some fault, accident or disaster with no backup available, then the data it is protecting may also be inaccessible / lost.
Lack of audit logging
If the key lifecycle is not fully recorded or logged, it will be more difficult to identify when a compromise has happened and any subsequent forensic investigation will be hampered.
Manual key management processes
The use of manual key management processes, using paper or inappropriate tools such as spreadsheets and accompanied by manual key ceremonies, can easily result in human errors that often go unnoticed and may leave keys highly vulnerable.
Mitigating the threats
So, what can be done to counter these threats and keep your keys (and your company) safe?
The only effective way to mitigate these threats is to use a dedicated electronic key management system, ideally a mature, proven solution from a reputable provider with good customer references. Any such key management system should utilize a hardware security module (HSM) to generate and protect keys, and to underpin the security of the whole system. If well-designed, such a system will offer the following benefits:
- Full lifecycle management of keys
- Generation of strong keys using a FIPS-certified RNG and hardware entropy source
- Protection of keys using a tamper-resistant HSM
- Strict policy-based controls to prevent the misuse/reuse of keys
- Automatic key rotation
- Automatic secure key distribution
- The ability to securely import/export keys in components or under a transport key
- The ability to securely destroy keys at the end of their lifecycle
- Strong user authentication, segregation of duties, and dual control over critical operations
- Intuitive user interface and secure workflow management to minimize the risk of human error
- Support for high-availability and business continuity
- Tamper-evident audit log, usage log and key histories for demonstrating compliance
- Ability to respond quickly to any detected compromise
Not only will such a system help protect your keys, it will also boost efficiency, reduce reliance on highly-skilled personnel, and simplify achieving, maintaining and demonstrating compliance with a multitude of standards and regulations such as GDPR, PCI-DSS, HIPAA, SOX and ISO 27001.
The biggest danger of all …
.. is inaction! The impact of a key compromise can be substantial:
- Forensic investigation costs
- Remediation costs
- Loss of sensitive information (e.g. industry secrets)
- Loss of competitive advantage
- Direct financial losses (e.g. illegitimate financial transactions)
- Compensation to customers
- Loss of reputation
- Loss of business
- Reduction in share price
- Dismissed executives
- Business closing down (as has been the result of some other data breaches)
Duty of reasonable care
An interesting court case in the USA as long ago as 1932, T.J. Hooper v. Northern Barge Corp., established that a company still has a reasonable duty of care towards using available technology, even where such technology may not be regarded as industry standard.
A company operates two tugs, each towing three barges full of coal for delivery. En route, the tugs encountered a storm which sank the last barge of each tug’s tow. The evidence suggests that there was a weather report broadcast over radio which would have warned the tug-captains of the weather and persuaded them to put into harbor. However, the tug-captains only had private radio receiving sets which were broken and their employer did not furnish them with sets for work. At the time of the incident, there was no industry standard or custom of furnishing all boats with radio receivers. [source]
The ruling concluded that “There are precautions so imperative that even their universal disregard will not excuse their omission … We hold the tugs therefore because had they been properly equipped, they would have got the Arlington reports. The injury was a direct consequence of this unseaworthiness.”
If we translate that into today’s world of key management, in the event of a legal case resulting from a key being compromised, a court may well find that if the defendant wasn’t using a key management system, a readily-available technology that could have prevented the incident, even though the use of such key management systems may not be considered industry-standard, then the defendant could be held to not be exercising a reasonable duty of care. The moral is that it is better to be seaworthy than to capsize through a lack of reasonable care!
- Key Size (retrieved 2018), Wikipedia
- NIST Special Publication 800-90B “Recommendation for the Entropy Sources Used for Random Bit Generation” (2018), by Meltem Sönmez Turan, Elaine Barker, John Kelsey, Kerry A. McKay, Mary L. Baish, Mike Boyle, National Institute of Standards and Technology
- Random Number Generator Attack, section “Prominent Examples” (retrieved 2018), Wikipedia
- On the Practical (In-)Security of 64-bit Block Ciphers (2016), by Karthikeyan Bhargavan, Gaëtan Leurent
- The Heartbleed Bug (2018), by Synopsis Inc.
- Flip Feng Shui vulnerability (2018), Systems and Security Group, VU Amsterdam
- Meltdown & Spectre – What you Need to Know about Protecting your Keys (2018), Rob Stubbs
- Meltdown and Spectre – Vulnerabilities in modern computers leak passwords and sensitive data (2018), MeltdownAttack
- The Radio-less Industry Standard Case (1932 / 2018), by Jonathan Zittrain, edited and republished by: Shailin Thomas, T.J. Hooper v. Northern Barge Corp.: https://h2o.law.harvard.edu/collages/4968
- Other Cryptomathic blog articles relating to Key Management: https://www.cryptomathic.com/news-events/blog/topic/key-management
Ever since cloud computing first emerged, security has been a prime concern of end users. The idea of handing over control of IT systems to third party operators by running hosted applications and infrastructure on remote servers has always sat uncomfortably with a significant number of business owners and CTOs.
These concerns are doubled when it comes to migrating the most sensitive data and the most mission critical applications into the cloud.
Over the years, cloud service providers have been able to win over most doubters on security with cutting edge privacy and anti-malware protections combined with state-of-the-art redundancy protocols. But on data, the disquiet persists. Out of your control, hosted on a remote multi-tenant server, who’s to say who really has access to it?
All data stored on a cloud platform, and indeed all data traffic communicated back and forth between the client and the host, is protected by encryption. But encryption is only as secure as the encryption keys used to decipher it. The question of ownership of the keys then raises its head. If they are also hosted in the same cloud system, managed by the same third-party operator, you are back to square one with the security concerns. If an external party can see or get hold of your encryption keys, they can get access to your data.
Bring Your Own Key (BYOK) is a protocol that aims to resolve this problem by maintaining a fundamental separation of encrypted data and encryption key. While your data might be entrusted into the hands of a cloud service provider, the key is not – at least, not in a way that forfeits the end user’s control over it, or makes it in any way accessible to external parties.
What does a fundamental separation of encryption and encryption key mean in practice? In the most simple terms, it means that your cloud host does not encrypt your data for you, or have anything to do with generating the key. BYOK means that encryption keys are generated, stored and applied completely independently by the client – they literally ‘bring their own key’ to enable encryption.
Taking HSM into the Cloud
To achieve secure encryption of data assets in an on-premise system, enterprises have long relied on hardware security modules (HSMs). An HSM is a cryptographic device which generates, stores and safeguards strong keys for the management of encryption across an IT system. Sophisticated and tamper resistant, an HSM can manage multiple digital keys for different use cases.
While HSMs remain a powerful cryptographic tool for on-premise systems, they were not designed for external encryption. Once you start introducing the public and private cloud, multiclouds and hybrid infrastructures into the equation, HSMs cease to be as effective. Until recently, there has been no accepted standard by which cloud providers will accept HSM-generated keys to work on their systems.
The main alternative to HSMs for cloud encryption to date has been Key Management Services (KMS). KMS is an encryption service offered and managed by a cloud provider for use on their own platforms. It offers all the key functionality you would get from an on-premise HSM, but solves the issue of compatibility. But the big drawback goes right to the heart of concerns over ownership and control of encryption – it hands the keys to the same people hosting your data services, which requires a considerable level of trust and removes any claim of ‘sole control’.
BYOK can be seen as offering a middle ground between HSM and KMS, some of the control but less of the overhead. Indeed, HSMs are integral to the BYOK concept. The approach has been pioneered by nCipher, which has developed an HSM product – nShield – that is compatible with the three biggest global cloud platforms, Microsoft Azure, Amazon Web Services (AWS) and Google Cloud.
nShield operates like any other HSM. Encryption keys are generated and stored within the on-premise device, meaning you are not handing over control to a third party provider – you are in charge of how and when the keys are used. The main difference is, nShield keys can be exported to the cloud service provider allowing the provider to encrypt data for applications hosted in the cloud as well as those run on premise.
With AWS and Google Cloud, keys are ‘leased’ to the host temporarily to encrypt digital assets on their servers. After an allocated period, the keys are automatically destroyed so there is no risk of them sitting on remote servers indefinitely waiting to fall into the wrong hands. Decryption is handled on the client side. Because the client retains the master key on premise, they can simply export a copy to the cloud provider as and when needed to update or change their encryption.
Azure Key Vault
With Azure, things work differently. nCipher has worked with Microsoft who host a highly available platform of nShield HSMs in the Azure cloud, the Azure Key Vault, and provide a protocol for securely storing encryption keys in it. The key is used to manage data encryption in the cloud environment, without the provider (Microsoft in this case) having any access to the data. In this formulation, BYOK completely integrates the robust security and control offered by HSM with the flexibility of the cloud.
The Azure Key Vault essentially works by encrypting the encryption key. Using the Transparent Data Encryption (TDE) approach, it encrypts the Database Encryption Key (DEK) stored on the boot page of a database using an asymmetric key known as a TDE Protector.
The TDE Protector can be generated in the client’s own on-premise HSM and then exported to the Azure Key vault via a secure bridge. But in what is arguably the biggest innovation of the nCipher and Microsoft BYOK approach (the systems leverage of nShield SecurityWorld controls end-to-end), the Azure Key Vault can also generate the TDE Protector itself. This is not like a KMS where the service provider generates and manages the key – this is all still controlled by the customer, but using a secure cloud environment rather than their own HSM.
The key point here is that even when the TDE Protector is generated within the Azure Vault, the cloud provider does not see and cannot extract the key. This maintains the confidence of full control for the client without the need for a highly available on-site HSM. Moreover, the HSM is not required once the tenant key is in the Azure Vault. Data services might be managed in the Azure cloud, but there is still effective and complete separation from responsibility for encryption and security, with the client managing keys and permissions for the Azure Vault migrated keys themselves.
BYOK is therefore all about establishing and maintaining trust in cloud data security by keeping all control of the encryption in the hands of the data owner – as they would have if they were running their databases on site. What the Azure approach demonstrates is that this confidence can still maintained even when the security protocols themselves are migrated to the cloud.
nCipher shows how nShield BYOK can strengthen your cloud key management practices.
We are interested to know what data you are planning to protect? Complete the ID-3 Data Protection Questionnaire to find out how we can help you.
ID-3 is an Oxford based consultancy with 15 years of global industry experience working with mainly finance, legal and government departments… (read more… )
The Quorum Oxford Business Park North Garsington Road Oxford OX4 2JZ
+44 (0) 1865 594724