<?xml version='1.0' encoding='utf-8'?>
<!DOCTYPE rfc [
  <!ENTITY nbsp    "&#160;">
  <!ENTITY zwsp   "&#8203;">
  <!ENTITY nbhy   "&#8209;">
  <!ENTITY wj     "&#8288;">
]>
<?xml-stylesheet type="text/xsl" href="rfc2629.xslt" ?>
<!-- generated by https://github.com/cabo/kramdown-rfc version 1.7.29 (Ruby 3.3.8) -->
<rfc xmlns:xi="http://www.w3.org/2001/XInclude" ipr="trust200902" docName="draft-irtf-t2trg-taxonomy-manufacturer-anchors-14" category="info" submissionType="IRTF" tocInclude="true" sortRefs="true" symRefs="true" version="3">
  <!-- xml2rfc v2v3 conversion 3.31.0 -->
  <front>
    <title abbrev="manufacturer keys">A Taxonomy of operational security considerations for manufacturer installed keys and Trust Anchors</title>
    <seriesInfo name="Internet-Draft" value="draft-irtf-t2trg-taxonomy-manufacturer-anchors-14"/>
    <author initials="M." surname="Richardson" fullname="Michael Richardson">
      <organization>Sandelman Software Works</organization>
      <address>
        <email>mcr+ietf@sandelman.ca</email>
      </address>
    </author>
    <date year="2025" month="December" day="03"/>
    <workgroup>T2TRG Research Group</workgroup>
    <keyword>Internet-Draft</keyword>
    <abstract>
      <?line 202?>

<t>This document provides a taxonomy of methods used by manufacturers of silicon and devices
to secure private keys and public trust anchors.
This deals with two related activities: how trust anchors and private keys
are installed into devices during manufacturing, and how the related
manufacturer held private keys are secured against disclosure.</t>
      <t>This document does not evaluate the different mechanisms, but rather just
serves to name them in a consistent manner in order to aid in communication.</t>
    </abstract>
    <note removeInRFC="true">
      <name>About This Document</name>
      <t>
        Status information for this document may be found at <eref target="https://datatracker.ietf.org/doc/draft-irtf-t2trg-taxonomy-manufacturer-anchors/"/>.
      </t>
      <t>
        Discussion of this document takes place on the
        t2trg Research Group mailing list (<eref target="mailto:t2trg@irtf.org"/>),
        which is archived at <eref target="https://mailarchive.ietf.org/arch/browse/t2trg/"/>.
        Subscribe at <eref target="https://www.ietf.org/mailman/listinfo/t2trg/"/>.
      </t>
      <t>Source for this draft and an issue tracker can be found at
        <eref target="https://github.com/t2trg/draft-irtf-t2trg-taxonomy-manufacturer-anchors"/>.</t>
    </note>
  </front>
  <middle>
    <?line 214?>

<section anchor="introduction">
      <name>Introduction</name>
      <t>An increasing number of protocols derive a significant part of their security by using trust anchors <xref target="RFC4949"/> that are installed by manufacturers into a device during manufacturing.
Disclosure of the list of trust anchors does not usually cause a problem, but changing them in any way does.
This includes adding, replacing or deleting anchors.</t>
      <t>The document <xref target="RFC6024"/> deals with how trust anchors are managed in the device which uses them.
This document deals with how the PKI associated with such a trust anchor is managed.</t>
      <t>Many protocols also leverage manufacturer installed identities.
These identities are usually in the form of <xref target="ieee802-1AR"/> Initial Device Identity certificates (IDevID).
The identity has two components: a private key that must remain under the strict control of a trusted part of the device, and a public part (the certificate), which (ignoring, for the moment, personal privacy concerns) may be freely disclosed.</t>
      <t>There also situations where identities are tied up in the provision of symmetric shared secrets.
A common example is the SIM card <xref target="_3GPP.51.011"/>.
Provisioning of a physical SIM card is generally considered a one-touch operation.
A virtual SIM (an eSIM) could be factory provisioned.</t>
      <t>It is further not unusual for many devices (particularly smartphones) to also have one or more group identity keys.
This is used, for instance, in <xref target="fidotechnote"/> to make claims about being a particular model of phone.
The key pair that does this is loaded into large batches of phones for privacy reasons:
If a single key was used, then this would allow tracking.
The shared key pair can not be used to identify a specific device.</t>
      <t>The trust anchors are used for a variety of purposes.
For instance, trust anchors are used to verify:</t>
      <ul spacing="normal">
        <li>
          <t>the signature on a software update (as per <xref target="RFC9019"/>),</t>
        </li>
        <li>
          <t>a TLS Server Certificate, such as when setting up an HTTPS connection,</t>
        </li>
        <li>
          <t>the <xref target="RFC8366"/> format voucher that provides proof of an ownership change.</t>
        </li>
      </ul>
      <t>Device identity keys are used when performing enrollment requests (in <xref target="RFC8995"/>, and in some uses of <xref target="RFC9140"/>.)
The device identity certificate is also used to sign Evidence by an Attesting Environment (see <xref target="RFC9334"/>).</t>
      <t>These security artifacts are used to anchor other chains of information such as: an
Entity Attestation Token (EAT) <xref target="RFC9711"/> Claim as to the version of software/firmware running on a device <xref target="I-D.birkholz-suit-coswid-manifest"/>, an EAT claim about legitimate network activity (via <xref target="I-D.ietf-iotops-mud-rats"/>, or embedded in the IDevID in <xref target="RFC8520"/>).</t>
      <t>Known software versions lead directly to vendor/distributor signed Software Bill of Materials (SBOM), such as those described by <xref target="RFC9393"/> and the NTIA/SBOM work <xref target="ntiasbom"/> and CISQ/OMG SBOM work underway <xref target="cisqsbom"/>.</t>
      <t>In order to manage risks and assess vulnerabilities in a Supply Chain, it is necessary to determine a degree of trustworthiness in each device.
A device may mislead audit systems as to its provenance, about its software load or even about what kind of device it is (see <xref target="RFC7168"/> for a humorous example).</t>
      <t>In order to properly assess the security of a Supply Chain it is necessary to understand the kinds and severity of the threats which a device has been designed to resist.
To do this, it is necessary to understand the ways in which the different trust anchors and identities are initially provisioned, are protected, and are updated.</t>
      <t>To do this, this document details the different trust anchors (TrAnc) and identities (IDs) found in typical devices.
The privacy and integrity of the TrAncs and IDs is often provided by a different, superior artifact.
This relationship is examined.</t>
      <t>While many might desire to assign numerical values to different mitigation techniques in order to be able to rank them, this document does not attempt to do that, as there are too many other (mostly human) factors that would come into play.
Such an effort is more properly in the purview of a formal processes such as <xref target="ISO27001"/>.</t>
      <section anchor="terminology">
        <name>Terminology</name>
        <t>This document is not a standards track document, and it does not make use of formal requirements language.</t>
        <t>This document defines a number of hyphenated terms, and they are summarized here:</t>
        <dl>
          <dt>birth certificates:</dt>
          <dd>
            <t>another term for IDevID</t>
          </dd>
          <dt>device-generated:</dt>
          <dd>
            <t>a private or symmetric key that is generated on the device</t>
          </dd>
          <dt>infrastructure-generated:</dt>
          <dd>
            <t>a private or symmetric key that is generated by some system, likely
located at the factory that built the device</t>
          </dd>
          <dt>IDevID certificates:</dt>
          <dd>
            <t>an initial device identity certificate as per <xref target="ieee802-1AR"/></t>
          </dd>
          <dt>key executives:</dt>
          <dd>
            <t>the people who are entrusted with pieces (shares) of a split secret <xref target="splittingnumbers"/></t>
          </dd>
          <dt>pkilevel:</dt>
          <dd>
            <t>the number of Certification Authorities in a public key infrastructure, counting the root, or trust-anchor as the first level, and not counting the End-Entity certificates <xref target="pkilevel"/></t>
          </dd>
          <dt>manufacturer installed certificates:</dt>
          <dd>
            <t>or MIC, see idevid</t>
          </dd>
          <dt>mechanically-installed:</dt>
          <dd>
            <t>when a key or certificate is programmed into non-volatile storage by
an out-of-band mechanism such as JTAG <xref target="JTAG"/></t>
          </dd>
          <dt>mechanically-transferred:</dt>
          <dd>
            <t>when a key or certificate is transferred into a system via private interface, such as serial console, JTAG managed mailbox, or other physically private interface</t>
          </dd>
          <dt>network-transferred:</dt>
          <dd>
            <t>when a key or certificate is transferred into a system using a network interface which would be available after the device has shipped.  This applies even if the network is physically attached using a bed-of-nails <xref target="BedOfNails"/>.</t>
          </dd>
          <dt>device/infrastructure-co-generated:</dt>
          <dd>
            <t>when a private or symmetric key is derived from a secret previously synchronized between the silicon vendor and the factory using a common algorithm.</t>
          </dd>
          <dt>TPM:</dt>
          <dd>
            <t>a Trusted Platform Module, such as the set of devices standardized by the Trusted Computing Group (TCG). See <xref target="TPM20spec"/>.</t>
          </dd>
        </dl>
        <t>In addition, <xref target="keygen"/> introduces three primary private key generation techniques named <em>arbitrarily</em> after three vegetables (avocado, brocolli, and carrot) and two secondary ones named after two fruits (salak and sapodilla).
The two secondary ones refer to methods where a secure element is involved, and mnemonically start with the same letter: S.</t>
      </section>
    </section>
    <section anchor="applicability-model">
      <name>Applicability Model</name>
      <t>There is a wide variety of devices to which this analysis can apply (see <xref target="I-D.ietf-iotops-7228bis"/>.)
This document will use a J-group processor as a sample.
This class is sufficiently large to experience complex issues among multiple CPUs, packages and operating systems, but at the same time, small enough that this class is often deployed in single-purpose IoT-like uses.
Devices in this class often have Secure Enclaves, and can include silicon manufacturer controlled processors in the boot process.
Note that access to the secure enclave is often not available to system integrators.
A very common IoT platform, the Raspberry PI, has a secure enclave with the GPU, but access to it is not available.</t>
      <t>Almost all larger systems (servers, laptops, desktops) include a Baseboard Management Controller (BMC), which ranges from an M-Group Class 3 MCU, to a J-Group Class 10 CPU (see, for instance <xref target="openbmc"/> which uses a Linux kernel and system inside the BMC).
As the BMC usually has complete access to the main CPU's memory, I/O hardware and disk, the boot path security of such a system needs to be understood first as being about the security of the BMC.</t>
      <section anchor="a-reference-manufacturingboot-process">
        <name>A reference manufacturing/boot process</name>
        <t>In order to provide for immutability and privacy of the critical TrAnc and IDs, many CPU manufacturers will provide for some kind of private memory area which is only accessible when the CPU is in certain privileged states.
See the Terminology and Architecture sections of <xref target="RFC9397"/>, notably Trusted Execution Environment (TEE), Rich Execution Environment (REE), and Trusted Application Manager (TAM).</t>
        <t>The private memory that is important is usually non-volatile and rather small.
It may be located inside the CPU silicon die, or it may be located externally.
If the memory is external, then it is usually encrypted by a hardware mechanism on the CPU, with only the key kept inside the CPU.</t>
        <t>The entire mechanism may be external to the CPU in the form of a hardware-TPM module, or it may be entirely internal to the CPU in the form of a firmware-TPM.
It may use a custom interface to the rest of the system, or it may implement the TPM 1.2 or TPM 2.0 specifications (see <xref target="TPM20spec"/>.)
Those details are important to performing a full evaluation, but do not matter much to this model (see initial-enclave-location below).</t>
        <t>During the manufacturing process, once the components have been soldered to the board, the system is usually put through a system-level test.
This is often done as a "bed-of-nails" test <xref target="BedOfNails"/>, where the board has key points attached mechanically to a test system.
A <xref target="JTAG"/> process tests the System Under Test, and then initializes some firmware into the still empty flash storage.</t>
        <t>It is now common for a factory test image to be loaded first: this image will include code to initialize the private memory key described above, and will include a first-stage bootloader and some kind of (primitive) Trusted Application Manager (TAM).
(The TAM is a piece of software that lives within the trusted execution environment.)</t>
        <t>Embedded in the first-stage bootloader will be a Trust Anchor that is able to verify the second-stage bootloader image.</t>
        <t>After the system has undergone testing, the factory test image is erased, leaving the first-stage bootloader.
One or more second-stage bootloader images are installed.
The production image may be installed at that time, or if the second-stage bootloader is able to install it over the network, it may be done that way instead.</t>
        <t>There are many variations of the above process, and this section is not attempting to be prescriptive, but to provide enough illustration to motivate subsequent terminology.</t>
        <t>The process may be entirely automated, or it may be entirely driven by humans working in the factory, or a combination of the above.</t>
        <t>These steps may all occur on an access-controlled assembly line, or the system boards may be shipped from one place to another (maybe another country) before undergoing testing.</t>
        <t>Some systems are intended to be shipped in a tamper-proof state, but it is usually not desirable that bed-of-nails testing be possible without tampering, so the initialization process is usually done prior to rendering the system tamper-proof.
An example of a one-way tamper-proof, weather resistant treatment might to mount the system board in a case and fill the case with resin.</t>
        <t>Quality control testing may be done prior to, as well as after, the application of tamper-proofing, as systems which do not pass inspection may be reworked to fix flaws, and this should ideally be impossible once the system has been made tamper-proof.</t>
      </section>
    </section>
    <section anchor="types-of-trust-anchors">
      <name>Types of Trust Anchors</name>
      <t>Trust Anchors are fundamentally public keys with authorizations implicitly attached through the code that references them.</t>
      <t>They are used to validate other digitally signed artifacts.
Typically, these are chains of PKIX certificates leading to an End-Entity certificate (EE).</t>
      <t>The chains are usually presented as part of an externally provided object, with the term "externally" to be understood as being as close as untrusted flash, to as far as objects retrieved over a network.</t>
      <t>There is no requirement that there be any chain at all: the trust anchor can be used to validate a signature over a target object directly.</t>
      <t>The trust anchors are often stored in the form of self-signed certificates.
The self-signature does not offer any cryptographic assurance, but it does provide a form of error detection, providing verification against non-malicious forms of data corruption.
If storage is at a premium (such as inside-CPU non-volatile storage) then only the public key itself need to be stored.
For a 256-bit ECDSA key, this is 32 bytes of space.</t>
      <t>The following subsections explain a number of different kinds of trust anchors that a manufacturer can include into a device.
Each kind of trust anchor has different kinds of trust associated with it based what authorization is associated with the key.
That is, these different anchors do different things, and so the risk associated with the anchor depends upon what those things that can be done are.</t>
      <t>There are some characteristics associated with each trust anchor that depend upon the associated risk.</t>
      <ol spacing="normal" type="1"><li>
          <t>There is a risk around whether or not a trust anchor can be replaced or modified so that it's a different anchor, (with the associated private key) controlled by a different entity.  This is not a risk associated with loss of the original manufacturer's private key, but with integrity of the public key.</t>
        </li>
        <li>
          <t>There is a risk that multiple trust anchors could be added to the device.  This would not replace the manufacturer's anchor, but augment it with additional trust anchors not authorized by the manufacturer.</t>
        </li>
        <li>
          <t>There is a risk that a trust anchor could be removed from the device, or rendered unusable. For instance, it might be impractical to replace a trust anchor, but it might be possible via fault injection or high-energy radiation for an attacker to corrupt one or two bytes of a trust anchor, rendering the anchor useless.</t>
        </li>
        <li>
          <t>There is a risk associated with compromise of the associate private key that goes with the (public key) of the trust anchor.  While this is the most obvious concern expressed: that an attacker gets control of the key, the above three risks may be more practical for some attackers.</t>
        </li>
        <li>
          <t>There is a risk that the manufacturer will lose access to the private key, without the private key being compromised.  No attacker possesses the private key, but neither can the manufacturer use the key anymore.</t>
        </li>
      </ol>
      <t>Should any of the above things occur, the manufacturer will be in a position where they might have to deploy new trust anchors to every device.
For some trust anchors, they may be safely replaced with an over-the-air update.
For the trust anchor that authorizes over-the-air updates, the manufacturer might be in a position where every single device has to be recalled and serviced with specialized hardware capable of updating the firmware.
For some devices, the hardware will have to be entirely replaced, typically at great cost.</t>
      <t>This has already happened, see for example: <xref target="leakedmsikey"/></t>
      <section anchor="first-stage-bootloader-trust-anchor">
        <name>First-Stage Bootloader Trust Anchor</name>
        <t>This anchor is part of the first-stage bootloader, and it is used to validate a second-stage bootloader which may be stored in external flash.</t>
        <t>This is called the first-stage bootloader trust anchor.
In most cases this trust anchor is burnt into fuses in the CPU, often within the die along with the first-stage bootloader itself.
It can not be changed without replacement of the entire device.
Replacement, removal or modification of this trust anchor is improbable.</t>
        <t>The anchor could be rendered unusable if additional fuses can be blown.
Some fuse implementations allow for bits to be changed from 1 (unblown), to 0 (blown).
Access to blow the fuses is usually restricted.
On some devices, it requires voltages not normally present, making this impossible to do by software.
However, it might be possible for an attacker to blow fuses using an external high voltage probe or via injection of gamma ray.   The likely result however is that the device would no longer boot.</t>
      </section>
      <section anchor="software-update-trust-anchor">
        <name>Software Update Trust Anchor</name>
        <t>This anchor is used to validate the main application (or operating system) load for the device.</t>
        <t>It can be stored in a number of places.
First, it may be identical to the First Bootloader Trust Anchor.</t>
        <t>Second, it may be stored in the second-stage bootloader, and therefore its integrity is protected by the First Bootloader Trust Anchor.</t>
        <t>Third, it may be stored in the application code itself, where the application validates updates to the application directly (update in place), or via a double-buffer arrangement.
The initial (factory) load of the application code initializes the trust arrangement.</t>
        <t>In this situation the application code is not started in a secured boot path.
The second-stage bootloader does not validate the application/operating system before starting it, but it may still provide measured boot mechanism (see <xref target="measuredboot"/>, section 2.3.)</t>
      </section>
      <section anchor="trusted-application-manager-anchor">
        <name>Trusted Application Manager Anchor</name>
        <t>This anchor is the key used by the <xref target="RFC9397"/> Trusted Application Manager (TAM).
Code which is signed by this anchor will be given execution privileges as described by the manifest which accompanies the code.</t>
        <t>The TAM software <em>itself</em> will typically be verified by the first-stage bootloader, so it needs to be signed by the First Bootloader Trust Anchor.
The TAM could also be a component of the first or second stage bootloader.</t>
        <t>This privilege may include updating anchors that are present within the TAM, or elsewhere in the Trusted Execution Environment.</t>
      </section>
      <section anchor="public-webpki-anchors">
        <name>Public WebPKI Anchors</name>
        <t>These anchors are used to authenticate HTTPS certificates from web sites.
These anchors are typically distributed as part of desktop browsers, and via desktop operating systems.
On IoT devices with specific purposes <xref target="RFC8520"/>, a limited number of connections is expected, so a limited number of trust anchors is usually distributed.</t>
        <t>The exact set of these anchors is not precisely defined: it is usually determined by the browser vendor (e.g., Mozilla, Google, Apple, Microsoft), or the operating system vendor (e.g., Apple, Google, Microsoft, Ubuntu).
In most cases these vendors look to the CA/Browser Forum <xref target="CABFORUM"/> for inclusion criteria.</t>
      </section>
      <section anchor="dnssec-root">
        <name>DNSSEC root</name>
        <t>This anchor is part of the DNS Security extensions.
It provides an anchor for integrity protection of DNS lookups.
Secure DNS lookups may be important in order to get access to software updates.
This anchor is now scheduled to change approximately every 3 years, with the new key announced several years before it is used, making it possible to embed keys that will be valid for up to five years.</t>
        <t>This trust anchor is typically part of the application/operating system code and is usually updated by the manufacturer when they do updates.
However, a system that is connected to the Internet may update the DNSSEC anchor itself through the mechanism described in <xref target="RFC5011"/>.</t>
        <t>There are concerns that there may be a chicken-and-egg situation for devices that have remained in a powered off state (or disconnected from the Internet) for a period of many years.
Upon being reconnected, the device would be unable to do DNSSEC validation because the trust anchor within the device would be too old.</t>
        <t>This could be fixed with a software/firmware update.
However, it could also be that in such a situation, that the device would be unable to validate the DNSSEC names required in order to obtain the operating system update.</t>
      </section>
      <section anchor="privatecloud-pki-anchors">
        <name>Private/Cloud PKI anchors</name>
        <t>It is common for many IoT and network appliances to have links to vendor provided services.
For instance, the IoT device that calls home for control purposes, or the network appliance that needs to validate a license key before it can operate.
(This may be identical to, or distinct from a Software Update anchor.  In particular, the device might call home over HTTPS to learn if there is a software update that needs to be done, but the update is signed by another key.)</t>
        <t>Such vendor services can be provided with public certificates, but the very short lifetime for public certificates <xref target="CABForum90day"/> precludes their use in many operational environments.
Instead, a private PKI anchor is included.
This can be in the form a multi-level PKI (as described in <xref target="pkilevel"/>), or degenerate to a level-1 PKI: a self-signed certificate.</t>
        <t>A level-1 PKI is very simple to create and operate, and there are innumerable situations where there is just a call to the "curl" utility <xref target="curl"/>, with a "--pinnedpubkey" option to specify the anchor.</t>
      </section>
      <section anchor="onboarding-and-other-enrollment-anchors">
        <name>Onboarding and other Enrollment anchors</name>
        <t><xref target="RFC8995"/>, <xref target="RFC8572"/>, and <xref target="RFC8366"/> specify mechanisms for onboarding of new devices.
The voucher artifact is transferred to the device by various means, and the device must verify the signature on it.
This requires a trust anchor to be built-in to the device, and some kind of private PKI be maintained by the vendor (or its authorized signing designate).
<xref target="I-D.ietf-anima-masa-considerations"/> provides some advice on choices in PKI design for a MASA.
The taxonomy presented in this document applies to describing how this PKI has been designed.</t>
      </section>
      <section anchor="onboarded-network-local-anchors">
        <name>Onboarded network-local anchors</name>
        <t><xref target="RFC7030"/>, <xref target="RFC8995"/> and <xref target="RFC9641"/> provide mechanisms by which new trust anchors may be loaded by a device during an onboarding process.
The trust anchors involved are typically local to an enterprise and are used to validate connections to other devices in the network.</t>
        <t>This typically includes connections to network management systems that may also load or modify other trust anchors in the system.
<xref target="I-D.ietf-anima-masa-considerations"/> provides some advice in the BRSKI <xref target="RFC8995"/> case for appropriate PKI complexity for such local PKIs.</t>
        <t>Note that this trust anchor is that of the network operator, is not related to the trust anchor described in the previous section that is used to validate an ownership transfer.</t>
      </section>
    </section>
    <section anchor="types-of-device-identities">
      <name>Types of Device Identities</name>
      <t>Device identities are installed during manufacturing time for a variety of purposes.</t>
      <t>Identities require some private component.
Asymmetric identities (e.g., RSA, ECDSA, EdDSA systems) require a corresponding public component, usually in the form of a certificate signed by a trusted third party.</t>
      <t>This certificate associates the identity with attributes.</t>
      <t>The process of making this coordinated key pair and then installing it into the device is called identity provisioning.</t>
      <section anchor="keygen">
        <name>Manufacturer installed IDevID certificates</name>
        <t><xref target="ieee802-1AR"/> defines a category of certificates that are installed by the manufacturer which contain a device unique serial number.</t>
        <t>A number of protocols depend upon this certificate.</t>
        <ul spacing="normal">
          <li>
            <t><xref target="RFC8572"/> and <xref target="RFC8995"/> introduce mechanisms for new devices (called pledges) to be onboarded into a network without intervention from an operator. A number of derived protocols such as <xref target="RFC9733"/>, <xref target="I-D.ietf-anima-constrained-voucher"/>, <xref target="I-D.ietf-anima-brski-cloud"/> extend this in a number of ways.</t>
          </li>
          <li>
            <t><xref target="RFC9334"/> depends upon a key provisioned into the Attesting Environment to sign Evidence. The IDevID may be used for this.</t>
          </li>
          <li>
            <t><xref target="RFC9019"/> depends upon a key provisioned into the
device in order to decrypt software updates.
Both symmetric and asymmetric keys are possible.
In both cases, the decrypt operation depends upon the device having access to
a private key provisioned in advance.
The IDevID can be used for this if algorithm choices permit.
ECDSA keys do not directly support encryption in the same way that RSA does, for
instance, but the addition of HPKE <xref target="RFC9180"/> allows for use of other key types.
There may be other legal considerations why the IDevID might not be used, and
a second key provisioned.</t>
          </li>
        </ul>
        <t>The manufacturer has the responsibility to provision a key pair into each
device as part of the manufacturing process.
There are a variety of mechanisms to accomplish this, which this section details.</t>
        <t>There are three fundamental ways to generate IDevID certificates for devices:</t>
        <ol spacing="normal" type="1"><li>
            <t>generating a private key on the device, creating a Certificate Signing
Request (or equivalent), and then returning a certificate to the device.</t>
          </li>
          <li>
            <t>generating a private key outside the device, signing the certificate, and
the installing both into the device.</t>
          </li>
          <li>
            <t>deriving the private key from a previously installed secret seed, that is shared with only the manufacturer.</t>
          </li>
        </ol>
        <t>There are additional variations where the IDevID is provided as part of a Trusted Platform Module (TPM) or Secure Element (SE).
The vendor may purchase such devices from another vendor, and that vendor often offers provisioning of a key pair into the device as a service.</t>
        <t>The document <xref target="I-D.moskowitz-ecdsa-pki"/> provides some practical instructions
on setting up a reference implementation for ECDSA keys using a three-tier
mechanism.</t>
        <t>The names of the five methods below are intended to be mnemonic.</t>
        <ul spacing="normal">
          <li>
            <t>Avocados have big seeds inside of them.</t>
          </li>
          <li>
            <t>Broccoli has a seed on the top (which is eaten).</t>
          </li>
          <li>
            <t>Carrots have a complicated seed process involving multiple years and many leaves.</t>
          </li>
          <li>
            <t>The two methods involving Secure Elements are named with the letter S.</t>
          </li>
        </ul>
        <section anchor="avcado">
          <name>Avocado method: On-device private key generation</name>
          <t>In this method, the device generates a private key on the device.
This is done within some very secure aspect of the device, such as in a Trusted Execution Environment, and the resulting private key is then stored in a secure and permanent way.
The permanency may extend beyond use of on-CPU flash, and could even involve blowing of one-time fuses.</t>
          <t>Generating the key on-device has the advantage that the private key never leaves the device.
The disadvantage is that the device may not have a verifiable random number generator.
The use of a pseudo-random number generator needs to be well seeded as explained in <xref target="RFC4086"/>.
<xref target="factoringrsa"/> is an example of a successful attack on this scenario.</t>
          <t>There are a number of options of how to get the public key from the device to the certification authority.
As it is a public key, privacy is less of a concern, and the focus is on integrity.
(However disclosing the public key may have impacts on trackability of the device.)</t>
          <t>So, transmission must be done in an integral manner, and must be securely associated with an assigned serial number.
The serial number goes into the certificate, and the resulting certificate needs to be loaded into the manufacturer's asset database, and returned to the device to be stored as the IDevID certificate.</t>
          <t>One way to do the transmission is during a factory Bed of Nails test (see <xref target="BedOfNails"/>) or JTAG Boundary Scan.
When done via a physical connection like this, then this is referred to as
<em>avocado device-generated</em> / <em>mechanically-transferred</em> method.</t>
          <t>There are other ways that could be used where a certificate signing request is sent over a special network channel when the device is powered up in the factory.
This is referred to as the <em>avocado device-generated</em> / <em>network-transferred</em> method.</t>
          <t>Regardless of how the certificate signing request is sent from the device to the factory, and how the certificate is returned to the device, a concern from production line managers is that the assembly line may have to wait for the certification authority to respond with the certificate.
This is inherently a synchronous process, as the process can not start until the private key is generated and stored.</t>
          <t>After the key generation, the device needs to set a flag such that it no longer will generate a new key, and will not accept a new IDevID via the factory network connection.
This may be a software setting, or could involve blowing a one-time fuse.</t>
          <t>Devices are typically constructed in a fashion such that the device is unable to ever disclose the private key via an external interface.
This is usually done using a secure-enclave provided by the CPU architecture in combination with on-chip non-volatile memory.</t>
          <t>The risk is that if an attacker with physical access is able to put the device back into an unconfigured mode, then the attacker may be able to substitute a new certificate into the device.
It is difficult to construct a rationale for doing this as the attacker would not be able to forge a certificate from the manufacturers' CA.
Other parties that rely on the IDevID would see the device as an imposter if another CA was used.
However, if the goal is theft of the device itself (without regard to having access to firmware updates), then use of another manufacturer identity may be profitable.
Stealing a very low value item, such as a light bulb makes very little sense.
Stealing medium value items, such as appliances, or high-value items such as cars, yachts or even airplanes would make sense.
Replacing the manufacturer IDevID permits the attacker to also replace the authority to transfer ownership in protocols like <xref target="RFC8995"/>.</t>
        </section>
        <section anchor="broccoli">
          <name>Broccoli method: Off-device private key generation</name>
          <t>In this method, a key pair is generated in the factory, outside of the device.
The factory keeps the private key in a restricted area, but uses it to form a Certification Signing Request (CSR).
The CSR is passed to the manufacturer's Certification Authority (CA), and a certificate is returned.
Other meta-data is often also returned, such as a serial number.</t>
          <t>Generating the key off-device has the advantage that the randomness of the private key can be better ensured.
As the private key is available to the manufacturing infrastructure, the authenticity of the public key is well known ahead of time.</t>
          <t>The private key and certificate can be programmed into the device along with the initial bootloader firmware in a single step.</t>
          <t>As the private key can be known to the factory in advance of the device being ready for it, the certificate can also be generated in advance.
This hides the latency to talk to the CA, and allows for the connectivity to the CA to be less reliable without shutting down the assembly line.
A single write to the flash of the device can contain the entire firmware of the device, including configuration of trust anchors and private keys.</t>
          <t>The major downside to generating the private key off-device is that it could be seen by the manufacturing infrastructure.
It could be compromised by humans in the factory, or the equipment could be compromised.
The use of this method increases the value of attacking the manufacturing infrastructure.</t>
          <t>If private keys are generated by the manufacturing plant, and are immediately installed, but never stored, then the window in which an attacker can gain access to the private key is immensely reduced.
But, the process then becomes more synchronous, negating much of the advantage of such a system.</t>
          <t>As in the previous case, the transfer may be done via physical interfaces such as bed-of-nails, giving the <em>brocolli infrastructure-generated</em> / <em>mechanically-transferred</em> method.</t>
          <t>There is also the possibility of having a <em>brocolli infrastructure-generated</em> / <em>network-transferred</em>
key.
There is a support for "server-generated" keys in <xref target="RFC7030"/>, <xref target="RFC8894"/>, and <xref target="RFC4210"/>.
All methods strongly recommend encrypting the private key for transfer.
This is difficult to comply with here as there is not yet any private key material in the device, so in many cases it will not be possible to encrypt the private key.
Still, it may be acceptable if the device is connected directly by a wired network and unroutable addresses are used.
A private wired network can often be made as secure as a bed-of-nails interface.</t>
        </section>
        <section anchor="carrot">
          <name>Carrot method: Key setup based on secret seed</name>
          <t>In this method, a random symmetric seed is generated by a component manufacturer.
This is typically the manufacturer of the CPU, often a system on a chip (SOC).
In this section there are two suppliers involved: the first is the familiar one that is responsible for the entire device (the device supplier), and the second one is the vendor of silicon chip.
This silicon chip is where a symmetric seed key has been provisioned.</t>
          <t>In this process, the silicon chip supplier provisions a unique secret into each device shipped.
This is typically at least 256-bits in size, but often is larger.</t>
          <t>This can be via fuses blown in a CPU's Trusted Execution Environment <xref target="RFC9397"/>, a TPM, or a Secure Element that is provisioned at its fabrication time.
In some cases, the secret is based upon a Physically Uncloneable Function <xref target="PUF"/>.</t>
          <t>This value is revealed to the board/system manufacturer only via a secure channel.
Upon first boot, the system (within a TEE, a TPM, or Secure Element) will generate a key pair using this seed to initialize a Pseudo-Random-Number-Generator (PRNG).
The manufacturer, in a separate system, will initialize the same PRNG and, thus generate the same key pair.
The manufacturer then derives the public key part, and signs it with their certification authority (CA), which inserts this public key into a certificate.
The private part then is destroyed by the manufacturer, ideally never stored or seen by anyone.
This same mechanism is used by the TCG for manufacturer provisioning of EK certificates in TPMs, as explained by Section 2.1.1 of <xref target="TPMEK"/>.</t>
          <t>The certificate (being public information) is placed into a database, in some cases it is loaded by the device as its IDevID certificate, in other cases, it is retrieved during the onboarding process based upon a unique serial number asserted by the device.</t>
          <t>In some ways, this method appears to have all the tradeoffs of the previous two methods: the device must correctly derive its own private key, and the manufacturer has access to the private key, making it also vulnerable.
But, the device does not depend upon any internal source of random numbers to derive its key.</t>
          <t>The manufacturer does all the private key manipulation in a secure place, probably offline, and need never involve the actual physical factory.
The manufacturer can even do this in a different country.</t>
          <t>The security of the process rests upon the difficulty in extracting the seed provided by the silicon chip supplier.
While the silicon chip supplier must operate their factory in a much more secure fashion, which has a much higher cost, the exposure for this facility can be much better controlled.
The device manufacturing factory, which has many more components as input, including device testing, can operate at a much lower risk level as a result.</t>
          <t>Additionally, there are some other advantages to the device manufacturer: The private keys and certificates may be calculated by the manufacturer asynchronously to the manufacturing process, either done in batches in advance of actual manufacturing, or on demand when an IDevID is requested.</t>
          <t>There are, however, additional downsides of this method for manufacturer: the cost of the silicon is often higher, and may involve additional discrete parts.
The higher cost is the result of outsourcing the security risk to the silicon manufacturer supplier.</t>
          <t>The weakest link in this process is that the resulting seeds must be communicated to the device manufacturer securely, usually in batches.
This will often by done by physical courier, but other arrangements are possible.
The device manufacturer must store and care for these seeds very carefully, but as soon as a certificate has been generated, the seeds can be destroyed.</t>
        </section>
        <section anchor="squash">
          <name>Squash method: on-device generation with Secure Element</name>
          <t>In this method, a key-pair is generated by the device using a secure element.
(It may be a discrete TPM, but when TPM is used to generate keys, that method is considered to fall into the <tt>avocado</tt> category)</t>
          <t>The secure element provides additional assurance that the private key was properly generated.
Secure elements are designed specifically so that private keys can not be extracted from the device, so even if the device is attacked in a sophisticated way, using hardware, the private key will not be disclosed.</t>
        </section>
        <section anchor="spinach">
          <name>Spinach method: Secure Element factory generation</name>
          <t>In this method, a key-pair is generated by a Secure Element silicon chip supplier in their factory.
This method is essentially identical to the squash method, but it occurs in a different factory!</t>
          <t>As a result the choice of which certification authority (CA) gets used may be very different.
It is typical for the silicon chip supplier to operate a CA themselves.
There are a few options:</t>
          <ol spacing="normal" type="1"><li>
              <t>they may put IDevIDs into the device which are generic to the silicon chip suppliere</t>
            </li>
            <li>
              <t>they may operate a CA on behalf of the device manufacturer,</t>
            </li>
            <li>
              <t>they may even connect over a network to the device manufacturer's CA.</t>
            </li>
          </ol>
          <t>The device manufacturer receives the secure element devices in batches in a similar way that they receive other parts.
The secure elements are placed by the device manufacturer's plant into the devices.</t>
          <t>Upon first boot the device manufacturer firmware can read the IDevID certificate that has been placed into the secure elements, and can ask the secure element to perform signing operations using the private key contained in the secure element.
But, the private key can not be extracted!</t>
          <t>Despite the increased convenience of this method, there may be a risk if the secure elements are stolen in transport.
A thief could use them to generate signatures that would appear to be from device manufacurer devices.
To deal with this, there is often a simple activation password that the device manufacturer firmware must provide to the secure element in order to activate it.
Not that this password is probably stored in the clear in the device manufacturer's firmware: it can't be encrypted, because the secure source of decryption keys would be in the secure element, creating a cyclic dependancy.</t>
          <t>Note that the Secure Element silicon vendor has to create an inventory specific for each client.</t>
        </section>
      </section>
    </section>
    <section anchor="public-key-infrastructures-pki">
      <name>Public Key Infrastructures (PKI)</name>
      <t>This section covers the attributes of the Public Key Infrastructure that a manufacturer must operate in order to provision IDevID certificates to devices.</t>
      <t><xref target="RFC5280"/> describes the format for PKIX certificates.
Numerous mechanisms for doing enrollment have been defined (including: EST <xref target="RFC7030"/>, CMP <xref target="RFC4210"/>, SCEP <xref target="RFC8894"/>).</t>
      <t><xref target="RFC5280"/> provides mechanisms to deal with multi-level certification
authorities, but it is not always clear what operating rules apply.</t>
      <t>The certification authority (CA) that is central to <xref target="RFC5280"/>-style public key infrastructures can suffer three kinds of failures:</t>
      <ol spacing="normal" type="1"><li>
          <t>disclosure of the private key,</t>
        </li>
        <li>
          <t>loss of access to the private key,</t>
        </li>
        <li>
          <t>inappropriate/unauthorized signing of a certificate or artifact by an unauthorized actor.</t>
        </li>
      </ol>
      <t>A PKI which discloses one or more private certification authority keys is no
longer secure.</t>
      <t>An attacker can create new identities, and forge certificates connecting
existing identities to attacker controlled public/private key pair.
This can permit the attacker to impersonate any specific device.</t>
      <t>In case 3, a failure occurs if the CA is convinced to sign (or issue) a certificate which it is not authorized to do so.
See for instance <xref target="ComodoGate"/>.
This is an authorization failure, and while a significant event, it does not result in the CA having to be re-initialized from scratch as no private keys were disclosed.</t>
      <t>This is distinguished from when access to the private key is lost, but the key has not been disclosed.
Any signatures that have already been made continue to be trustworthy, however, no new signature can be made.
This renders the CA useless.
If this the firmware signing key, it likely results in a recall of all products that need to use this trust anchor.
(Note: that there are some situations where a trust anchor will be created, it will then be used to sign a number of subordinate CAs, enough for the anticipated lifespan of the anchor, and then the private key intentionally destroyed.)</t>
      <t>If the PKI uses Certificate Revocation Lists (CRL), then an attacker that has access to the private key can also revoke existing identities.</t>
      <t>In the other direction, a PKI which loses access to a private key can no
longer function.
This does not immediately result in a failure, as existing identities remain valid until their expiry time (notAfter).
However, if CRLs or OCSP are in use, then the inability to sign a fresh CRL or OCSP response will result in all identities becoming invalid once the existing CRLs or OCSP statements expire.</t>
      <t>This section details some nomenclature about the structure of certification
authorities.</t>
      <section anchor="pkilevel">
        <name>Number of levels of certification authorities (pkilevel)</name>
        <t><xref section="6.1" sectionFormat="of" target="RFC5280"/> provides a Basic Path Validation.
In the formula, the certificates are arranged into a list.</t>
        <t>The certification authority (CA) starts with a Trust Anchor.
This is counted as the first level of the authority.</t>
        <t>In the degenerate case of a self-signed certificate, this a one level PKI.</t>
        <artset>
          <artwork type="svg"><svg xmlns="http://www.w3.org/2000/svg" version="1.1" height="112" width="128" viewBox="0 0 128 112" class="diagram" text-anchor="middle" font-family="monospace" font-size="13px">
              <path d="M 8,32 L 8,96" fill="none" stroke="black"/>
              <path d="M 96,32 L 96,96" fill="none" stroke="black"/>
              <path d="M 120,48 L 120,80" fill="none" stroke="black"/>
              <path d="M 8,32 L 96,32" fill="none" stroke="black"/>
              <path d="M 104,80 L 120,80" fill="none" stroke="black"/>
              <path d="M 8,96 L 96,96" fill="none" stroke="black"/>
              <g class="text">
                <text x="108" y="52">&lt;-</text>
                <text x="40" y="68">Issuer=</text>
                <text x="80" y="68">X</text>
                <text x="48" y="84">Subject=X</text>
              </g>
            </svg>
          </artwork>
          <artwork type="ascii-art"><![CDATA[
.----------.
|          |<-.
|Issuer= X |  |
|Subject=X |--'
'----------'
]]></artwork>
        </artset>
        <t>The private key associated with the Trust Anchor signs one or more certificates.
When this first level authority signs only End-Entity (EE) certificates,
then this is a two level PKI.</t>
        <artset>
          <artwork type="svg"><svg xmlns="http://www.w3.org/2000/svg" version="1.1" height="272" width="232" viewBox="0 0 232 272" class="diagram" text-anchor="middle" font-family="monospace" font-size="13px">
              <path d="M 8,32 L 8,96" fill="none" stroke="black"/>
              <path d="M 8,176 L 8,224" fill="none" stroke="black"/>
              <path d="M 32,96 L 32,168" fill="none" stroke="black"/>
              <path d="M 80,96 L 80,128" fill="none" stroke="black"/>
              <path d="M 96,32 L 96,96" fill="none" stroke="black"/>
              <path d="M 96,176 L 96,224" fill="none" stroke="black"/>
              <path d="M 120,48 L 120,80" fill="none" stroke="black"/>
              <path d="M 136,176 L 136,224" fill="none" stroke="black"/>
              <path d="M 144,128 L 144,168" fill="none" stroke="black"/>
              <path d="M 224,176 L 224,224" fill="none" stroke="black"/>
              <path d="M 8,32 L 96,32" fill="none" stroke="black"/>
              <path d="M 96,80 L 120,80" fill="none" stroke="black"/>
              <path d="M 8,96 L 96,96" fill="none" stroke="black"/>
              <path d="M 80,128 L 144,128" fill="none" stroke="black"/>
              <path d="M 8,176 L 40,176" fill="none" stroke="black"/>
              <path d="M 64,176 L 96,176" fill="none" stroke="black"/>
              <path d="M 136,176 L 168,176" fill="none" stroke="black"/>
              <path d="M 192,176 L 224,176" fill="none" stroke="black"/>
              <path d="M 8,224 L 96,224" fill="none" stroke="black"/>
              <path d="M 136,224 L 224,224" fill="none" stroke="black"/>
              <polygon class="arrowhead" points="152,168 140,162.4 140,173.6" fill="black" transform="rotate(90,144,168)"/>
              <polygon class="arrowhead" points="40,168 28,162.4 28,173.6" fill="black" transform="rotate(90,32,168)"/>
              <g class="text">
                <text x="108" y="52">&lt;-</text>
                <text x="40" y="68">Issuer=</text>
                <text x="80" y="68">X</text>
                <text x="164" y="68">root</text>
                <text x="48" y="84">Subject=X</text>
                <text x="164" y="84">CA</text>
                <text x="52" y="180">EE</text>
                <text x="180" y="180">EE</text>
                <text x="40" y="196">Issuer=</text>
                <text x="80" y="196">X</text>
                <text x="168" y="196">Issuer=</text>
                <text x="208" y="196">X</text>
                <text x="52" y="212">Subject=Y1</text>
                <text x="180" y="212">Subject=Y2</text>
              </g>
            </svg>
          </artwork>
          <artwork type="ascii-art"><![CDATA[
.----------.
|          |<-.
|Issuer= X |  |   root
|Subject=X +--'    CA
'--+-----+-'
   |     |
   |     '-------.
   |             |
   v             v
.----EE----.    .----EE----.
|Issuer= X |    |Issuer= X |
|Subject=Y1|    |Subject=Y2|
'----------'    '----------'


]]></artwork>
        </artset>
        <t>When this first level authority signs subordinate certification authorities,
and those certification authorities sign End-Entity certificates, then
this is a three level PKI.</t>
        <artset>
          <artwork type="svg"><svg xmlns="http://www.w3.org/2000/svg" version="1.1" height="448" width="480" viewBox="0 0 480 448" class="diagram" text-anchor="middle" font-family="monospace" font-size="13px">
              <path d="M 8,304 L 8,352" fill="none" stroke="black"/>
              <path d="M 32,176 L 32,224" fill="none" stroke="black"/>
              <path d="M 32,256 L 32,296" fill="none" stroke="black"/>
              <path d="M 56,128 L 56,168" fill="none" stroke="black"/>
              <path d="M 56,224 L 56,256" fill="none" stroke="black"/>
              <path d="M 88,224 L 88,256" fill="none" stroke="black"/>
              <path d="M 96,304 L 96,352" fill="none" stroke="black"/>
              <path d="M 120,176 L 120,224" fill="none" stroke="black"/>
              <path d="M 128,32 L 128,96" fill="none" stroke="black"/>
              <path d="M 136,304 L 136,352" fill="none" stroke="black"/>
              <path d="M 152,96 L 152,128" fill="none" stroke="black"/>
              <path d="M 152,256 L 152,296" fill="none" stroke="black"/>
              <path d="M 200,96 L 200,128" fill="none" stroke="black"/>
              <path d="M 216,32 L 216,96" fill="none" stroke="black"/>
              <path d="M 224,304 L 224,352" fill="none" stroke="black"/>
              <path d="M 240,48 L 240,80" fill="none" stroke="black"/>
              <path d="M 256,304 L 256,352" fill="none" stroke="black"/>
              <path d="M 272,256 L 272,296" fill="none" stroke="black"/>
              <path d="M 280,176 L 280,224" fill="none" stroke="black"/>
              <path d="M 304,128 L 304,168" fill="none" stroke="black"/>
              <path d="M 304,224 L 304,256" fill="none" stroke="black"/>
              <path d="M 344,224 L 344,256" fill="none" stroke="black"/>
              <path d="M 344,304 L 344,352" fill="none" stroke="black"/>
              <path d="M 368,176 L 368,224" fill="none" stroke="black"/>
              <path d="M 384,304 L 384,352" fill="none" stroke="black"/>
              <path d="M 400,256 L 400,296" fill="none" stroke="black"/>
              <path d="M 472,304 L 472,352" fill="none" stroke="black"/>
              <path d="M 128,32 L 216,32" fill="none" stroke="black"/>
              <path d="M 216,80 L 240,80" fill="none" stroke="black"/>
              <path d="M 128,96 L 216,96" fill="none" stroke="black"/>
              <path d="M 56,128 L 152,128" fill="none" stroke="black"/>
              <path d="M 200,128 L 304,128" fill="none" stroke="black"/>
              <path d="M 32,176 L 120,176" fill="none" stroke="black"/>
              <path d="M 280,176 L 368,176" fill="none" stroke="black"/>
              <path d="M 32,224 L 120,224" fill="none" stroke="black"/>
              <path d="M 280,224 L 368,224" fill="none" stroke="black"/>
              <path d="M 32,256 L 56,256" fill="none" stroke="black"/>
              <path d="M 88,256 L 152,256" fill="none" stroke="black"/>
              <path d="M 272,256 L 304,256" fill="none" stroke="black"/>
              <path d="M 344,256 L 400,256" fill="none" stroke="black"/>
              <path d="M 8,304 L 40,304" fill="none" stroke="black"/>
              <path d="M 64,304 L 96,304" fill="none" stroke="black"/>
              <path d="M 136,304 L 168,304" fill="none" stroke="black"/>
              <path d="M 192,304 L 224,304" fill="none" stroke="black"/>
              <path d="M 256,304 L 288,304" fill="none" stroke="black"/>
              <path d="M 312,304 L 344,304" fill="none" stroke="black"/>
              <path d="M 384,304 L 416,304" fill="none" stroke="black"/>
              <path d="M 440,304 L 472,304" fill="none" stroke="black"/>
              <path d="M 8,352 L 96,352" fill="none" stroke="black"/>
              <path d="M 136,352 L 224,352" fill="none" stroke="black"/>
              <path d="M 256,352 L 344,352" fill="none" stroke="black"/>
              <path d="M 384,352 L 472,352" fill="none" stroke="black"/>
              <polygon class="arrowhead" points="408,296 396,290.4 396,301.6" fill="black" transform="rotate(90,400,296)"/>
              <polygon class="arrowhead" points="312,168 300,162.4 300,173.6" fill="black" transform="rotate(90,304,168)"/>
              <polygon class="arrowhead" points="280,296 268,290.4 268,301.6" fill="black" transform="rotate(90,272,296)"/>
              <polygon class="arrowhead" points="160,296 148,290.4 148,301.6" fill="black" transform="rotate(90,152,296)"/>
              <polygon class="arrowhead" points="64,168 52,162.4 52,173.6" fill="black" transform="rotate(90,56,168)"/>
              <polygon class="arrowhead" points="40,296 28,290.4 28,301.6" fill="black" transform="rotate(90,32,296)"/>
              <g class="text">
                <text x="228" y="52">&lt;-</text>
                <text x="44" y="68">root</text>
                <text x="160" y="68">Issuer=</text>
                <text x="200" y="68">X</text>
                <text x="44" y="84">CA</text>
                <text x="168" y="84">Subject=X</text>
                <text x="64" y="196">Issuer=</text>
                <text x="104" y="196">X</text>
                <text x="200" y="196">subordinate</text>
                <text x="312" y="196">Issuer=</text>
                <text x="352" y="196">X</text>
                <text x="76" y="212">Subject=Y1</text>
                <text x="196" y="212">CA</text>
                <text x="324" y="212">Subject=Y2</text>
                <text x="52" y="308">EE</text>
                <text x="180" y="308">EE</text>
                <text x="300" y="308">EE</text>
                <text x="428" y="308">EE</text>
                <text x="40" y="324">Issuer=</text>
                <text x="84" y="324">Y1</text>
                <text x="168" y="324">Issuer=</text>
                <text x="212" y="324">Y1</text>
                <text x="288" y="324">Issuer=</text>
                <text x="332" y="324">Y2</text>
                <text x="416" y="324">Issuer=</text>
                <text x="460" y="324">Y2</text>
                <text x="52" y="340">Subject=Z1</text>
                <text x="180" y="340">Subject=Z1</text>
                <text x="300" y="340">Subject=Z3</text>
                <text x="428" y="340">Subject=Z4</text>
              </g>
            </svg>
          </artwork>
          <artwork type="ascii-art"><![CDATA[
               .----------.
               |          |<-.
   root        |Issuer= X |  |
    CA         |Subject=X +--'
               '--+-----+-'
                  |     |
      .-----------'     '------------.
      |                              |
      v                              v
   .----------.                   .----------.
   |Issuer= X |    subordinate    |Issuer= X |
   |Subject=Y1|        CA         |Subject=Y2|
   '--+---+---'                   '--+----+--'
      |   |                          |    |
   .--'   '-------.              .---'    '------.
   |              |              |               |
   v              v              v               v
.----EE----.    .----EE----.   .----EE----.    .----EE----.
|Issuer= Y1|    |Issuer= Y1|   |Issuer= Y2|    |Issuer= Y2|
|Subject=Z1|    |Subject=Z1|   |Subject=Z3|    |Subject=Z4|
'----------'    '----------'   '----------'    '----------'





]]></artwork>
        </artset>
        <t>In general, when arranged as a tree, with the End-Entity certificates at the
bottom, and the Trust Anchor at the top, then the pkilevel is defined to be where the deepest EE
certificates are, counting from one.</t>
        <t>It is quite common to have a three-level PKI, where the root (level one) of the CA is
stored in a Hardware Security Module (HSM) in a way that it cannot be continuously accessed ("offline"), while the level two subordinate CA can sign certificates at any time ("online").</t>
      </section>
      <section anchor="protection-of-ca-private-keys">
        <name>Protection of CA private keys</name>
        <t>The private key for the certification authorities must be protected from
disclosure.
The strongest protection is afforded by keeping them in an offline device,
passing Certificate Signing Requests (CSRs) to the offline device by human
process.</t>
        <t>For examples of extreme measures, see <xref target="kskceremony"/>.
There is however a wide spectrum of needs, as exampled in <xref target="rootkeyceremony"/>.
The SAS70 audit standard is usually used as a basis for the Ceremony, see <xref target="keyceremony2"/>.</t>
        <t>This is inconvenient, and may involve latencies of days, possibly even weeks
to months if the offline device is kept in a locked environment that requires
multiple keys to be present.</t>
        <t>There is therefore a tension between protection and availability.
Convenient and timely access to sign new artifacts is not something that is just nice to have.
If access is inconvenient then it may cause delays for signing of new code releases,
or it may incentivize technical staff to build in workarounds in order that they can get their job done faster.
The compromise between situations is often mitigated by having some levels of the PKI be offline, and some levels of the PKI be online.</t>
      </section>
      <section anchor="preservation-of-ca-and-trust-anchor-private-keys">
        <name>Preservation of CA and Trust Anchor private keys</name>
        <t>A public key (or certificate) is installed into target device(s) as a trust anchor.
It is there in order to verify further artifacts, and it represents a significant investment.
Trust anchors must not be easily replaced by attackers, and securing the trust anchor against such tampering may involve burning the trust anchor into unchangeable fuses inside a CPU.</t>
        <t>Replacement of the anchor can involve a physical recall of every single device.
It is therefore important that the trust anchor is usable for the entire lifetime of every single one of the devices.</t>
        <t>The previous section deals with attacks against the infrastructure: the attacker wants to get access to the private key material, or to convince the infrastructure to use the private key material to their bidding.
Such an event, if it was undetected would be catastrophic.
It would render almost every device useless (or potentially dangerous) until the anchor could be replaced.</t>
        <t>There is a different situation, however, which would lead to a similar result.
If the legitimate owner of the trust anchor infrastructure loses access to the private keys, then an equally catastrophic situation occurs.</t>
        <t>There are many situations that could lead to this.
The most typical situation would seem to be some kind of physical damage: a flood, a fire.
Less obvious situations could occur if a human forgets a password, or if the human with the password(s) dies, or becomes incapacitated.</t>
        <t>Backups of critical material are routinely done.
Storage of backups offsite deals with physical damage, and in many cases the organization maintains an entire set of equipment at another location.</t>
        <t>The question then becomes: how are the backups unlocked, or activated.
Why attack the primary site physically if an attacker can target the backup site, or the people whose job it is to activate the backup site?</t>
        <t>Consider the situation where a hurricane or earthquake takes out all power and communications at an organizations' primary location, and it becomes necessary to activate the backup site.
What does it take to do that?</t>
        <t>Typically, the secrets will be split using <xref target="shamir79"/> into multiple pieces, each piece being carried with a different trusted employee.</t>
        <t>In <xref target="kskceremony"/>, the pieces are stored on smartcards that are kept in a vault, and the trusted people carry keys to the vault.</t>
        <t>One advantage of this mechanism is that if necessary, the doors to the vault can be drilled out.
This takes some significant time and leaves significant evidence, so it can not be done quietly by an attacker.
In the case of the DNSSEC Root, a failure of the vault to open actually required this to be done.</t>
        <t>In other systems the digital pieces are carried on the person themselves, ideally encrypted with a password known only to that person.</t>
        <t><xref target="shamir79"/> allows for keys to be split up into many components (n of them), where only some smaller number of them, k, need to be present to reconstruct the secret.
This is known as a (k, n) threshold scheme.</t>
        <section anchor="splittingnumbers">
          <name>Secret splitting, k-of-n</name>
          <t>In this document, each of the people who hold a piece of the secret are referred to as Key Executives.</t>
          <t>The choice of n, and the choice of k is therefore of critical concern.
It seems unwise for an organization to publish them, as it provides some evidence as to how many Key Executives would need to be coerced.</t>
          <t>The identities of the n Key Executives should also be confidential.
The list of who they are should probably be limited to the members of the board and executive.
There does not seem to be any particular reason for the Key Executives to be members of the board, but having a long term relationship with the enterprise seems reasonable, and a clear understanding of when to use the piece.</t>
          <t>The number k, which is the minimum number of people that would need to be coerced should also remain confidential.</t>
          <t>A number that can be published is the difference between k and n, which represents the number of redundant Key Executives that exist.</t>
          <t>An enterprise that has operations in multiple places may be better positioned to survive incidents that disrupt travel.
For instance, an earthquake, tsunami, or pandemic not only has the possibility to kill Key Executives but it could also damage smartcards or USB key that they are stored on.
<xref target="shamir79"/> suggests that n=2k-1, which implies that a simple majority of Key Executives are needed to reconstruct the secret.</t>
          <t>However there are other values of k with some different tradeoffs: requiring a majority of Key Executives might be difficult when travel is impossible.
For instance, a value of k set to be less than a simple majority.
This enables the situation where Key Executives are split between two or more continents (with each continent having at least k Key Executives) would allow either continent to continue operations without the other group.</t>
          <t>This might be a very good way to manage a code signing or update signing key.
Split it among development groups in three time zones (eight hours apart), such that any of those development groups can issue an emergency security patch.
Another way would be to have three End-Entity certificates that can sign code, and have each time zone sign their own code.  That implies that there is at least a level two PKI around the code signing process, and that any bootloaders that need to verify the code being starting it are able to do PKI operations.</t>
        </section>
      </section>
      <section anchor="supporting-provisioned-anchors-in-devices">
        <name>Supporting provisioned anchors in devices</name>
        <t>IDevID-type Identity (or Birth) Certificates which are provisioned into
devices need to be signed by a certification authority maintained by the manufacturer.
During the period of manufacture of new product, the manufacturer needs to be able to sign new Identity Certificates.</t>
        <t>During the anticipated lifespan of the devices the manufacturer needs to maintain the ability for third parties to validate the Identity Certificates.
If there are Certificate Revocation Lists (CRLs) involved, then they will need to re-signed during this period.
Even for devices with a short active lifetime, the lifespan of the device could be very long if devices are kept in a warehouse for many decades before being activated.</t>
        <t>Trust anchors which are provisioned in the devices will have corresponding
private keys maintained by the manufacturer.
The trust anchors will often anchor a PKI which is going to be used for a
particular purpose.
There will be End-Entity (EE) certificates of this PKI which will be used to sign
particular artifacts (such as software updates), or messages in communications protocols
(such as TLS connections).
The private keys associated with these EE certificates are not stored in the
device, but are maintained by the manufacturer.
These need even more care than the private keys stored in the devices, as
compromise of the software update key compromises all the devices, not
just a single device.</t>
      </section>
    </section>
    <section anchor="evaluation-questions">
      <name>Evaluation Questions</name>
      <t>This section recaps the set of questions that may need to be answered.
This document does not assign valuation to the answers.</t>
      <section anchor="integrity-and-confidentiality-of-on-device-data">
        <name>Integrity and Confidentiality of on-device data</name>
        <dl>
          <dt>initial-enclave-location:</dt>
          <dd>
            <t>Is the location of the initial software trust anchor internal to the CPU package?
Some systems have a software verification public key which is built into the CPU package, while other systems store that initial key in a non-volatile device external to the CPU.</t>
          </dd>
          <dt>initial-enclave-integrity-key:</dt>
          <dd>
            <t>If the first-stage bootloader is external to the CPU, it will often need integrity protection. To verify that, another key is needed.  It could be a keyed hash.   Where is the key for this hash located?</t>
          </dd>
          <dt>initial-enclave-confidentiality-key:</dt>
          <dd>
            <t>If the first-stage data is external to the CPU, is it kept confidential by use of encryption?</t>
          </dd>
          <dt>first-stage-exposure:</dt>
          <dd>
            <t>The number of people involved in the first stage initialization.
An entirely automated system would have a number zero.
A factory with three 8 hour shifts might have a number that is a multiple of three.
A system with humans involved may be subject to bribery attacks, while a system with no humans may be subject to attacks on the system which are hard to notice.</t>
          </dd>
          <dt>first-second-stage-gap:</dt>
          <dd>
            <t>how far and long does a board travel between being initialized with a first-stage bootloader to where it is locked down so that changes to the bootloader can no longer be made.
For many situations, there is no distance at all as they occur in the same factory, but for other situations boards are manufactured and tested in one location, but are initialized elsewhere.</t>
          </dd>
        </dl>
      </section>
      <section anchor="integrity-and-privacy-of-device-identity-infrastructure">
        <name>Integrity and Privacy of device identity infrastructure</name>
        <t>For IDevID provisioning, which includes a private key and matching
certificate installed into the device, the associated public key
infrastructure that anchors this identity must be maintained by the
manufacturer.</t>
        <dl>
          <dt>identity-pki-level:</dt>
          <dd>
            <t>referring to <xref target="pkilevel"/>, the level number at which End-Entity certificates are present.</t>
          </dd>
          <dt>identity-time-limits-per-subordinate:</dt>
          <dd>
            <t>how long is each subordinate CA maintained before a new
subordinate CA key is generated?  There may be no time limit, only a device
count limit.</t>
          </dd>
          <dt>identity-number-per-subordinate:</dt>
          <dd>
            <t>how many identities are signed by a particular subordinate CA before it is
retired?  There may be no numeric limit, only a time limit.</t>
          </dd>
          <dt>identity-anchor-storage:</dt>
          <dd>
            <t>how is the root CA key stored? An open description that might include whether an HSM is used, or not, or even the model of it.</t>
          </dd>
          <dt>identity-shared-split-extra:</dt>
          <dd>
            <t>referring to <xref target="splittingnumbers"/>, where a private key is split up into n-components, of which k are required to recover the key, this number is equal to n-k.
This is the number of spare shares.
Publishing this provides a measure of how much redundancy is present while not actually revealing either k or n.</t>
          </dd>
          <dt>identity-shared-split-continents:</dt>
          <dd>
            <t>the number of continents on which the private key can be recovered without trans-continental travel by any of the secret shareholders.</t>
          </dd>
        </dl>
      </section>
      <section anchor="integrity-and-privacy-of-included-trust-anchors">
        <name>Integrity and Privacy of included trust anchors</name>
        <t>For each trust anchor (public key) stored in the device, there will be an
associated PKI.
For each of those PKI the following questions need to be answered.</t>
        <dl>
          <dt>pki-level:</dt>
          <dd>
            <t>how deep is the EE that will be evaluated, based upon the criteria in <xref target="pkilevel"/></t>
          </dd>
          <dt>pki-algorithms:</dt>
          <dd>
            <t>what kind of algorithms and key sizes can actively be used with the device.</t>
          </dd>
          <dt>pki-lifespan:</dt>
          <dd>
            <t>what is the timespan for this anchor.  Does it get replaced at some interval, and if so, by what means is this done?</t>
          </dd>
          <dt>pki-level-locked:</dt>
          <dd>
            <t>(a Boolean) the level where the EE cert will be found locked by the device, or can
levels be added or deleted by the PKI operator without code changes to the
device.</t>
          </dd>
          <dt>pki-breadth:</dt>
          <dd>
            <t>how many different non-expired EE certificates is the PKI designed to manage?</t>
          </dd>
          <dt>pki-lock-policy:</dt>
          <dd>
            <t>can any EE certificate be used with this trust anchor to sign?  Or, is there
some kind of policy OID or Subject restriction?  Are specific subordinate
CAs needed that lead to the EE?</t>
          </dd>
          <dt>pki-anchor-storage:</dt>
          <dd>
            <t>how is the private key associated with this trust anchor stored? How many people are needed to recover it?</t>
          </dd>
        </dl>
      </section>
    </section>
    <section anchor="security-considerations">
      <name>Security Considerations</name>
      <t>This entire document is about security considerations.</t>
    </section>
    <section anchor="iana-considerations">
      <name>IANA Considerations</name>
      <t>This document makes no IANA requests.</t>
    </section>
    <section anchor="acknowledgements">
      <name>Acknowledgements</name>
      <t>Robert Martin of MITRE provided some guidance about citing the SBOM efforts.
Carsten Bormann provides many editorial suggestions.</t>
    </section>
    <section anchor="changelog">
      <name>Changelog</name>
    </section>
  </middle>
  <back>
    <references anchor="sec-combined-references">
      <name>References</name>
      <references anchor="sec-normative-references">
        <name>Normative References</name>
        <reference anchor="RFC5280">
          <front>
            <title>Internet X.509 Public Key Infrastructure Certificate and Certificate Revocation List (CRL) Profile</title>
            <author fullname="D. Cooper" initials="D." surname="Cooper"/>
            <author fullname="S. Santesson" initials="S." surname="Santesson"/>
            <author fullname="S. Farrell" initials="S." surname="Farrell"/>
            <author fullname="S. Boeyen" initials="S." surname="Boeyen"/>
            <author fullname="R. Housley" initials="R." surname="Housley"/>
            <author fullname="W. Polk" initials="W." surname="Polk"/>
            <date month="May" year="2008"/>
            <abstract>
              <t>This memo profiles the X.509 v3 certificate and X.509 v2 certificate revocation list (CRL) for use in the Internet. An overview of this approach and model is provided as an introduction. The X.509 v3 certificate format is described in detail, with additional information regarding the format and semantics of Internet name forms. Standard certificate extensions are described and two Internet-specific extensions are defined. A set of required certificate extensions is specified. The X.509 v2 CRL format is described in detail along with standard and Internet-specific extensions. An algorithm for X.509 certification path validation is described. An ASN.1 module and examples are provided in the appendices. [STANDARDS-TRACK]</t>
            </abstract>
          </front>
          <seriesInfo name="RFC" value="5280"/>
          <seriesInfo name="DOI" value="10.17487/RFC5280"/>
        </reference>
        <reference anchor="ieee802-1AR" target="http://standards.ieee.org/findstds/standard/802.1AR-2009.html">
          <front>
            <title>IEEE 802.1AR Secure Device Identifier</title>
            <author>
              <organization>IEEE Standard</organization>
            </author>
            <date year="2009"/>
          </front>
        </reference>
      </references>
      <references anchor="sec-informative-references">
        <name>Informative References</name>
        <reference anchor="RFC8995">
          <front>
            <title>Bootstrapping Remote Secure Key Infrastructure (BRSKI)</title>
            <author fullname="M. Pritikin" initials="M." surname="Pritikin"/>
            <author fullname="M. Richardson" initials="M." surname="Richardson"/>
            <author fullname="T. Eckert" initials="T." surname="Eckert"/>
            <author fullname="M. Behringer" initials="M." surname="Behringer"/>
            <author fullname="K. Watsen" initials="K." surname="Watsen"/>
            <date month="May" year="2021"/>
            <abstract>
              <t>This document specifies automated bootstrapping of an Autonomic Control Plane. To do this, a Secure Key Infrastructure is bootstrapped. This is done using manufacturer-installed X.509 certificates, in combination with a manufacturer's authorizing service, both online and offline. We call this process the Bootstrapping Remote Secure Key Infrastructure (BRSKI) protocol. Bootstrapping a new device can occur when using a routable address and a cloud service, only link-local connectivity, or limited/disconnected networks. Support for deployment models with less stringent security requirements is included. Bootstrapping is complete when the cryptographic identity of the new key infrastructure is successfully deployed to the device. The established secure connection can be used to deploy a locally issued certificate to the device as well.</t>
            </abstract>
          </front>
          <seriesInfo name="RFC" value="8995"/>
          <seriesInfo name="DOI" value="10.17487/RFC8995"/>
        </reference>
        <reference anchor="RFC9019">
          <front>
            <title>A Firmware Update Architecture for Internet of Things</title>
            <author fullname="B. Moran" initials="B." surname="Moran"/>
            <author fullname="H. Tschofenig" initials="H." surname="Tschofenig"/>
            <author fullname="D. Brown" initials="D." surname="Brown"/>
            <author fullname="M. Meriac" initials="M." surname="Meriac"/>
            <date month="April" year="2021"/>
            <abstract>
              <t>Vulnerabilities in Internet of Things (IoT) devices have raised the need for a reliable and secure firmware update mechanism suitable for devices with resource constraints. Incorporating such an update mechanism is a fundamental requirement for fixing vulnerabilities, but it also enables other important capabilities such as updating configuration settings and adding new functionality.</t>
              <t>In addition to the definition of terminology and an architecture, this document provides the motivation for the standardization of a manifest format as a transport-agnostic means for describing and protecting firmware updates.</t>
            </abstract>
          </front>
          <seriesInfo name="RFC" value="9019"/>
          <seriesInfo name="DOI" value="10.17487/RFC9019"/>
        </reference>
        <reference anchor="RFC9140">
          <front>
            <title>Nimble Out-of-Band Authentication for EAP (EAP-NOOB)</title>
            <author fullname="T. Aura" initials="T." surname="Aura"/>
            <author fullname="M. Sethi" initials="M." surname="Sethi"/>
            <author fullname="A. Peltonen" initials="A." surname="Peltonen"/>
            <date month="December" year="2021"/>
            <abstract>
              <t>The Extensible Authentication Protocol (EAP) provides support for multiple authentication methods. This document defines the EAP-NOOB authentication method for nimble out-of-band (OOB) authentication and key derivation. The EAP method is intended for bootstrapping all kinds of Internet-of-Things (IoT) devices that have no preconfigured authentication credentials. The method makes use of a user-assisted, one-directional, out-of-band (OOB) message between the peer device and authentication server to authenticate the in-band key exchange. The device must have a nonnetwork input or output interface, such as a display, microphone, speaker, or blinking light, that can send or receive dynamically generated messages of tens of bytes in length.</t>
            </abstract>
          </front>
          <seriesInfo name="RFC" value="9140"/>
          <seriesInfo name="DOI" value="10.17487/RFC9140"/>
        </reference>
        <reference anchor="RFC9334">
          <front>
            <title>Remote ATtestation procedureS (RATS) Architecture</title>
            <author fullname="H. Birkholz" initials="H." surname="Birkholz"/>
            <author fullname="D. Thaler" initials="D." surname="Thaler"/>
            <author fullname="M. Richardson" initials="M." surname="Richardson"/>
            <author fullname="N. Smith" initials="N." surname="Smith"/>
            <author fullname="W. Pan" initials="W." surname="Pan"/>
            <date month="January" year="2023"/>
            <abstract>
              <t>In network protocol exchanges, it is often useful for one end of a communication to know whether the other end is in an intended operating state. This document provides an architectural overview of the entities involved that make such tests possible through the process of generating, conveying, and evaluating evidentiary Claims. It provides a model that is neutral toward processor architectures, the content of Claims, and protocols.</t>
            </abstract>
          </front>
          <seriesInfo name="RFC" value="9334"/>
          <seriesInfo name="DOI" value="10.17487/RFC9334"/>
        </reference>
        <reference anchor="RFC9393">
          <front>
            <title>Concise Software Identification Tags</title>
            <author fullname="H. Birkholz" initials="H." surname="Birkholz"/>
            <author fullname="J. Fitzgerald-McKay" initials="J." surname="Fitzgerald-McKay"/>
            <author fullname="C. Schmidt" initials="C." surname="Schmidt"/>
            <author fullname="D. Waltermire" initials="D." surname="Waltermire"/>
            <date month="June" year="2023"/>
            <abstract>
              <t>ISO/IEC 19770-2:2015 Software Identification (SWID) tags provide an extensible XML-based structure to identify and describe individual software components, patches, and installation bundles. SWID tag representations can be too large for devices with network and storage constraints. This document defines a concise representation of SWID tags: Concise SWID (CoSWID) tags. CoSWID supports a set of semantics and features that are similar to those for SWID tags, as well as new semantics that allow CoSWIDs to describe additional types of information, all in a more memory-efficient format.</t>
            </abstract>
          </front>
          <seriesInfo name="RFC" value="9393"/>
          <seriesInfo name="DOI" value="10.17487/RFC9393"/>
        </reference>
        <reference anchor="RFC9733">
          <front>
            <title>BRSKI with Alternative Enrollment (BRSKI-AE)</title>
            <author fullname="D. von Oheimb" initials="D." role="editor" surname="von Oheimb"/>
            <author fullname="S. Fries" initials="S." surname="Fries"/>
            <author fullname="H. Brockhaus" initials="H." surname="Brockhaus"/>
            <date month="March" year="2025"/>
            <abstract>
              <t>This document defines enhancements to the Bootstrapping Remote Secure Key Infrastructure (BRSKI) protocol, known as BRSKI with Alternative Enrollment (BRSKI-AE). BRSKI-AE extends BRSKI to support certificate enrollment mechanisms instead of the originally specified use of Enrollment over Secure Transport (EST). It supports certificate enrollment protocols such as the Certificate Management Protocol (CMP) that use authenticated self-contained signed objects for certification messages, allowing for flexibility in network device onboarding scenarios. The enhancements address use cases where the existing enrollment mechanism may not be feasible or optimal, providing a framework for integrating suitable alternative enrollment protocols. This document also updates the BRSKI reference architecture to accommodate these alternative methods, ensuring secure and scalable deployment across a range of network environments.</t>
            </abstract>
          </front>
          <seriesInfo name="RFC" value="9733"/>
          <seriesInfo name="DOI" value="10.17487/RFC9733"/>
        </reference>
        <reference anchor="I-D.ietf-anima-brski-cloud">
          <front>
            <title>Bootstrapping Remote Secure Key Infrastructure (BRSKI) Cloud Registrar</title>
            <author fullname="Owen Friel" initials="O." surname="Friel">
              <organization>Cisco</organization>
            </author>
            <author fullname="Rifaat Shekh-Yusef" initials="R." surname="Shekh-Yusef">
              <organization>Ciena</organization>
            </author>
            <author fullname="Michael Richardson" initials="M." surname="Richardson">
              <organization>Sandelman Software Works</organization>
            </author>
            <date day="9" month="September" year="2025"/>
            <abstract>
              <t>   Bootstrapping Remote Secure Key Infrastructures (BRSKI) defines how
   to onboard a device securely into an operator-maintained
   infrastructure.  It assumes that there is local network
   infrastructure for the device to discover.  On networks without that,
   there is nothing present to help onboard the device.

   This document extends BRSKI and defines behavior for bootstrapping
   devices for deployments where no local infrastructure is available,
   such as in a home or remote office.  This document defines how the
   device can use a well-defined "call-home" mechanism to find the
   operator-maintained infrastructure.

   This document defines how to contact a well-known Cloud Registrar,
   and two ways in which the device may be redirected towards the
   operator-maintained infrastructure.  The Cloud Registrar enables
   discovery of the operator-maintained infrastructure, and may enable
   establishment of trust with operator-maintained infrastructure that
   does not support BRSKI mechanisms.

   This document updates RFC 8995 (BRSKI).

              </t>
            </abstract>
          </front>
          <seriesInfo name="Internet-Draft" value="draft-ietf-anima-brski-cloud-19"/>
        </reference>
        <reference anchor="I-D.ietf-iotops-7228bis">
          <front>
            <title>Terminology for Constrained-Node Networks</title>
            <author fullname="Carsten Bormann" initials="C." surname="Bormann">
              <organization>Universität Bremen TZI</organization>
            </author>
            <author fullname="Mehmet Ersue" initials="M." surname="Ersue">
         </author>
            <author fullname="Ari Keränen" initials="A." surname="Keränen">
              <organization>Ericsson</organization>
            </author>
            <author fullname="Carles Gomez" initials="C." surname="Gomez">
              <organization>Universitat Politecnica de Catalunya</organization>
            </author>
            <date day="4" month="November" year="2025"/>
            <abstract>
              <t>   The Internet Protocol Suite is increasingly used on small devices
   with severe constraints on power, memory, and processing resources,
   creating constrained-node networks.  This document provides a number
   of basic terms that have been useful in research and standardization
   work for constrained-node networks.

   This document obsoletes RFC 7228.

              </t>
            </abstract>
          </front>
          <seriesInfo name="Internet-Draft" value="draft-ietf-iotops-7228bis-03"/>
        </reference>
        <reference anchor="I-D.ietf-anima-constrained-voucher">
          <front>
            <title>Constrained Bootstrapping Remote Secure Key Infrastructure (cBRSKI)</title>
            <author fullname="Michael Richardson" initials="M." surname="Richardson">
              <organization>Sandelman Software Works</organization>
            </author>
            <author fullname="Peter Van der Stok" initials="P." surname="Van der Stok">
              <organization>vanderstok consultancy</organization>
            </author>
            <author fullname="Panos Kampanakis" initials="P." surname="Kampanakis">
              <organization>Cisco Systems</organization>
            </author>
            <author fullname="Esko Dijk" initials="E." surname="Dijk">
              <organization>IoTconsultancy.nl</organization>
            </author>
            <date day="18" month="October" year="2025"/>
            <abstract>
              <t>   This document defines the Constrained Bootstrapping Remote Secure Key
   Infrastructure (cBRSKI) protocol, which provides a solution for
   secure zero-touch onboarding of resource-constrained (IoT) devices
   into the network of a domain owner.  This protocol is designed for
   constrained networks, which may have limited data throughput or may
   experience frequent packet loss. cBRSKI is a variant of the BRSKI
   protocol, which uses an artifact signed by the device manufacturer
   called the "voucher" which enables a new device and the owner's
   network to mutually authenticate.  While the BRSKI voucher data is
   encoded in JSON, cBRSKI uses a compact CBOR-encoded voucher.  The
   BRSKI voucher data definition is extended with new data types that
   allow for smaller voucher sizes.  The Enrollment over Secure
   Transport (EST) protocol, used in BRSKI, is replaced with EST-over-
   CoAPS; and HTTPS used in BRSKI is replaced with DTLS-secured CoAP
   (CoAPS).  This document Updates RFC 8995 and RFC 9148.

              </t>
            </abstract>
          </front>
          <seriesInfo name="Internet-Draft" value="draft-ietf-anima-constrained-voucher-29"/>
        </reference>
        <reference anchor="I-D.ietf-anima-masa-considerations">
          <front>
            <title>Operational Considerations for Voucher infrastructure for BRSKI MASA</title>
            <author fullname="Michael Richardson" initials="M." surname="Richardson">
              <organization>Sandelman Software Works</organization>
            </author>
            <author fullname="Wei Pan" initials="W." surname="Pan">
              <organization>Huawei Technologies</organization>
            </author>
            <date day="5" month="November" year="2025"/>
            <abstract>
              <t>   This document describes a number of operational modes that a BRSKI
   Manufacturer Authorized Signing Authority (MASA) may take on.

   Each mode is defined, and then each mode is given a relevance within
   an over applicability of what kind of organization the MASA is
   deployed into.  This document does not change any protocol
   mechanisms.

              </t>
            </abstract>
          </front>
          <seriesInfo name="Internet-Draft" value="draft-ietf-anima-masa-considerations-01"/>
        </reference>
        <reference anchor="I-D.moskowitz-ecdsa-pki">
          <front>
            <title>Guide for building an ECC pki</title>
            <author fullname="Robert Moskowitz" initials="R." surname="Moskowitz">
              <organization>HTT Consulting</organization>
            </author>
            <author fullname="Henk Birkholz" initials="H." surname="Birkholz">
              <organization>Fraunhofer SIT</organization>
            </author>
            <author fullname="Liang Xia" initials="L." surname="Xia">
              <organization>Huawei</organization>
            </author>
            <author fullname="Michael Richardson" initials="M." surname="Richardson">
              <organization>Sandelman Software Works</organization>
            </author>
            <date day="31" month="January" year="2021"/>
            <abstract>
              <t>   This memo provides a guide for building a PKI (Public Key
   Infrastructure) using openSSL.  All certificates in this guide are
   ECDSA, P-256, with SHA256 certificates.  Along with common End Entity
   certificates, this guide provides instructions for creating IEEE
   802.1AR iDevID Secure Device certificates.

              </t>
            </abstract>
          </front>
          <seriesInfo name="Internet-Draft" value="draft-moskowitz-ecdsa-pki-10"/>
        </reference>
        <reference anchor="RFC4949">
          <front>
            <title>Internet Security Glossary, Version 2</title>
            <author fullname="R. Shirey" initials="R." surname="Shirey"/>
            <date month="August" year="2007"/>
            <abstract>
              <t>This Glossary provides definitions, abbreviations, and explanations of terminology for information system security. The 334 pages of entries offer recommendations to improve the comprehensibility of written material that is generated in the Internet Standards Process (RFC 2026). The recommendations follow the principles that such writing should (a) use the same term or definition whenever the same concept is mentioned; (b) use terms in their plainest, dictionary sense; (c) use terms that are already well-established in open publications; and (d) avoid terms that either favor a particular vendor or favor a particular technology or mechanism over other, competing techniques that already exist or could be developed. This memo provides information for the Internet community.</t>
            </abstract>
          </front>
          <seriesInfo name="FYI" value="36"/>
          <seriesInfo name="RFC" value="4949"/>
          <seriesInfo name="DOI" value="10.17487/RFC4949"/>
        </reference>
        <reference anchor="RFC5011">
          <front>
            <title>Automated Updates of DNS Security (DNSSEC) Trust Anchors</title>
            <author fullname="M. StJohns" initials="M." surname="StJohns"/>
            <date month="September" year="2007"/>
            <abstract>
              <t>This document describes a means for automated, authenticated, and authorized updating of DNSSEC "trust anchors". The method provides protection against N-1 key compromises of N keys in the trust point key set. Based on the trust established by the presence of a current anchor, other anchors may be added at the same place in the hierarchy, and, ultimately, supplant the existing anchor(s).</t>
              <t>This mechanism will require changes to resolver management behavior (but not resolver resolution behavior), and the addition of a single flag bit to the DNSKEY record. [STANDARDS-TRACK]</t>
            </abstract>
          </front>
          <seriesInfo name="STD" value="74"/>
          <seriesInfo name="RFC" value="5011"/>
          <seriesInfo name="DOI" value="10.17487/RFC5011"/>
        </reference>
        <reference anchor="RFC8366">
          <front>
            <title>A Voucher Artifact for Bootstrapping Protocols</title>
            <author fullname="K. Watsen" initials="K." surname="Watsen"/>
            <author fullname="M. Richardson" initials="M." surname="Richardson"/>
            <author fullname="M. Pritikin" initials="M." surname="Pritikin"/>
            <author fullname="T. Eckert" initials="T." surname="Eckert"/>
            <date month="May" year="2018"/>
            <abstract>
              <t>This document defines a strategy to securely assign a pledge to an owner using an artifact signed, directly or indirectly, by the pledge's manufacturer. This artifact is known as a "voucher".</t>
              <t>This document defines an artifact format as a YANG-defined JSON document that has been signed using a Cryptographic Message Syntax (CMS) structure. Other YANG-derived formats are possible. The voucher artifact is normally generated by the pledge's manufacturer (i.e., the Manufacturer Authorized Signing Authority (MASA)).</t>
              <t>This document only defines the voucher artifact, leaving it to other documents to describe specialized protocols for accessing it.</t>
            </abstract>
          </front>
          <seriesInfo name="RFC" value="8366"/>
          <seriesInfo name="DOI" value="10.17487/RFC8366"/>
        </reference>
        <reference anchor="RFC8572">
          <front>
            <title>Secure Zero Touch Provisioning (SZTP)</title>
            <author fullname="K. Watsen" initials="K." surname="Watsen"/>
            <author fullname="I. Farrer" initials="I." surname="Farrer"/>
            <author fullname="M. Abrahamsson" initials="M." surname="Abrahamsson"/>
            <date month="April" year="2019"/>
            <abstract>
              <t>This document presents a technique to securely provision a networking device when it is booting in a factory-default state. Variations in the solution enable it to be used on both public and private networks. The provisioning steps are able to update the boot image, commit an initial configuration, and execute arbitrary scripts to address auxiliary needs. The updated device is subsequently able to establish secure connections with other systems. For instance, a device may establish NETCONF (RFC 6241) and/or RESTCONF (RFC 8040) connections with deployment-specific network management systems.</t>
            </abstract>
          </front>
          <seriesInfo name="RFC" value="8572"/>
          <seriesInfo name="DOI" value="10.17487/RFC8572"/>
        </reference>
        <reference anchor="RFC7030">
          <front>
            <title>Enrollment over Secure Transport</title>
            <author fullname="M. Pritikin" initials="M." role="editor" surname="Pritikin"/>
            <author fullname="P. Yee" initials="P." role="editor" surname="Yee"/>
            <author fullname="D. Harkins" initials="D." role="editor" surname="Harkins"/>
            <date month="October" year="2013"/>
            <abstract>
              <t>This document profiles certificate enrollment for clients using Certificate Management over CMS (CMC) messages over a secure transport. This profile, called Enrollment over Secure Transport (EST), describes a simple, yet functional, certificate management protocol targeting Public Key Infrastructure (PKI) clients that need to acquire client certificates and associated Certification Authority (CA) certificates. It also supports client-generated public/private key pairs as well as key pairs generated by the CA.</t>
            </abstract>
          </front>
          <seriesInfo name="RFC" value="7030"/>
          <seriesInfo name="DOI" value="10.17487/RFC7030"/>
        </reference>
        <reference anchor="RFC8894">
          <front>
            <title>Simple Certificate Enrolment Protocol</title>
            <author fullname="P. Gutmann" initials="P." surname="Gutmann"/>
            <date month="September" year="2020"/>
            <abstract>
              <t>This document specifies the Simple Certificate Enrolment Protocol (SCEP), a PKI protocol that leverages existing technology by using Cryptographic Message Syntax (CMS, formerly known as PKCS #7) and PKCS #10 over HTTP. SCEP is the evolution of the enrolment protocol sponsored by Cisco Systems, which enjoys wide support in both client and server implementations, as well as being relied upon by numerous other industry standards that work with certificates.</t>
            </abstract>
          </front>
          <seriesInfo name="RFC" value="8894"/>
          <seriesInfo name="DOI" value="10.17487/RFC8894"/>
        </reference>
        <reference anchor="RFC4210">
          <front>
            <title>Internet X.509 Public Key Infrastructure Certificate Management Protocol (CMP)</title>
            <author fullname="C. Adams" initials="C." surname="Adams"/>
            <author fullname="S. Farrell" initials="S." surname="Farrell"/>
            <author fullname="T. Kause" initials="T." surname="Kause"/>
            <author fullname="T. Mononen" initials="T." surname="Mononen"/>
            <date month="September" year="2005"/>
            <abstract>
              <t>This document describes the Internet X.509 Public Key Infrastructure (PKI) Certificate Management Protocol (CMP). Protocol messages are defined for X.509v3 certificate creation and management. CMP provides on-line interactions between PKI components, including an exchange between a Certification Authority (CA) and a client system. [STANDARDS-TRACK]</t>
            </abstract>
          </front>
          <seriesInfo name="RFC" value="4210"/>
          <seriesInfo name="DOI" value="10.17487/RFC4210"/>
        </reference>
        <reference anchor="_3GPP.51.011" target="http://www.3gpp.org/ftp/Specs/archive/51_series/51.011/51011-4f0.zip">
          <front>
            <title>Specification of the Subscriber Identity Module - Mobile Equipment (SIM-ME) interface</title>
            <author>
              <organization abbrev="3GPP">3rd Generation Partnership Project</organization>
              <address>
                <postal>
                  <country>France</country>
                  <city>Sophia Antipolis Cedex</city>
                </postal>
              </address>
            </author>
            <author fullname="PHAN, Ly Thanh">
              <organization>Gemalto N.V.</organization>
            </author>
            <date day="15" month="June" year="2005"/>
          </front>
        </reference>
        <reference anchor="RFC6024">
          <front>
            <title>Trust Anchor Management Requirements</title>
            <author fullname="R. Reddy" initials="R." surname="Reddy"/>
            <author fullname="C. Wallace" initials="C." surname="Wallace"/>
            <date month="October" year="2010"/>
            <abstract>
              <t>A trust anchor represents an authoritative entity via a public key and associated data. The public key is used to verify digital signatures, and the associated data is used to constrain the types of information for which the trust anchor is authoritative. A relying party uses trust anchors to determine if a digitally signed object is valid by verifying a digital signature using the trust anchor's public key, and by enforcing the constraints expressed in the associated data for the trust anchor. This document describes some of the problems associated with the lack of a standard trust anchor management mechanism and defines requirements for data formats and push-based protocols designed to address these problems. This document is not an Internet Standards Track specification; it is published for informational purposes.</t>
            </abstract>
          </front>
          <seriesInfo name="RFC" value="6024"/>
          <seriesInfo name="DOI" value="10.17487/RFC6024"/>
        </reference>
        <reference anchor="CABForum90day" target="   https://cabforum.org/working-groups/server/baseline-requirements/requirements/">
          <front>
            <title>CA/Browser Forum Latest Baseline Requirements, version 2.1.5</title>
            <author>
              <organization>CA/Browser Forum</organization>
            </author>
            <date>n.d.</date>
          </front>
        </reference>
        <reference anchor="factoringrsa" target="https://eprint.iacr.org/2013/599">
          <front>
            <title>Factoring RSA keys from certified smart cards: Coppersmith in the wild</title>
            <author initials="D. J." surname="Bernstein" fullname="Daniel J. Bernstein">
              <organization/>
            </author>
            <author initials="Y.-A." surname="Chang" fullname="Yun-An Chang">
              <organization/>
            </author>
            <author initials="C.-M." surname="Cheng" fullname="Chen-Mou Cheng">
              <organization/>
            </author>
            <author initials="L.-P." surname="Chou" fullname="Li-Ping Chou">
              <organization/>
            </author>
            <author initials="N." surname="Heninger" fullname="Nadia Heninger">
              <organization/>
            </author>
            <author initials="T." surname="Lange" fullname="Tanja Lange">
              <organization/>
            </author>
            <author initials="N. van" surname="Someren" fullname="Nicko van Someren">
              <organization/>
            </author>
            <date year="2013" month="September" day="16"/>
          </front>
        </reference>
        <reference anchor="kskceremony" target="https://www.iana.org/dnssec/dps/zsk-operator/dps-zsk-operator-v2.0.pdf">
          <front>
            <title>DNSSEC Practice Statement for the Root Zone ZSK Operator</title>
            <author>
              <organization>Verisign</organization>
            </author>
            <date year="2017"/>
          </front>
        </reference>
        <reference anchor="rootkeyceremony" target="https://cryptography.fandom.com/wiki/Root_Key_Ceremony">
          <front>
            <title>Root Key Ceremony, Cryptography Wiki</title>
            <author>
              <organization>Community</organization>
            </author>
            <date year="2020" month="April" day="04"/>
          </front>
        </reference>
        <reference anchor="keyceremony2" target="http://www.digi-sign.com/compliance/key%20ceremony/index">
          <front>
            <title>SAS 70 Key Ceremony</title>
            <author>
              <organization>Digi-Sign</organization>
            </author>
            <date year="2020" month="April" day="04"/>
          </front>
        </reference>
        <reference anchor="shamir79" target="http://web.mit.edu/6.857/OldStuff/Fall03/ref/Shamir-HowToShareASecret.pdf">
          <front>
            <title>How to share a secret.</title>
            <author initials="A." surname="Shamir" fullname="Adi Shamir">
              <organization/>
            </author>
            <date year="1979"/>
          </front>
        </reference>
        <reference anchor="fidotechnote" target="https://fidoalliance.org/fido-technotes-the-truth-about-attestation/">
          <front>
            <title>FIDO TechNotes: The Truth about Attestation</title>
            <author>
              <organization>FIDO Alliance</organization>
            </author>
            <date year="2018" month="July" day="19"/>
          </front>
        </reference>
        <reference anchor="ntiasbom" target="https://www.ntia.doc.gov/SoftwareTransparency">
          <front>
            <title>NTIA Software Component Transparency</title>
            <author>
              <organization>NTIA</organization>
            </author>
            <date year="2020" month="July" day="01"/>
          </front>
        </reference>
        <reference anchor="cisqsbom" target="https://www.it-cisq.org/software-bill-of-materials/index.htm">
          <front>
            <title>TOOL-TO-TOOL SOFTWARE BILL OF MATERIALS EXCHANGE</title>
            <author>
              <organization>CISQ/Object Management Group</organization>
            </author>
            <date year="2020" month="July" day="01"/>
          </front>
        </reference>
        <reference anchor="ComodoGate" target="https://www.theregister.com/2011/03/28/comodo_gate_hacker_breaks_cover/">
          <front>
            <title>Comodo-gate hacker brags about forged certificate exploit</title>
            <author>
              <organization/>
            </author>
            <date year="2011" month="March" day="28"/>
          </front>
        </reference>
        <reference anchor="openbmc" target="https://www.openbmc.org/">
          <front>
            <title>Defining a Standard Baseboard Management Controller Firmware Stack</title>
            <author>
              <organization>Linux Foundation/OpenBMC Group</organization>
            </author>
            <date year="2020" month="July" day="01"/>
          </front>
        </reference>
        <reference anchor="JTAG" target="https://en.wikipedia.org/wiki/JTAG">
          <front>
            <title>Joint Test Action Group</title>
            <author>
              <organization/>
            </author>
            <date year="2020" month="August" day="26"/>
          </front>
        </reference>
        <reference anchor="BedOfNails" target="https://en.wikipedia.org/wiki/Bed_of_nails_tester">
          <front>
            <title>Bed of nails tester</title>
            <author>
              <organization>Wikipedia</organization>
            </author>
            <date year="2024" month="December" day="30"/>
          </front>
        </reference>
        <reference anchor="leakedmsikey" target="https://www.binarly.io/blog/leaked-msi-source-code-with-intel-oem-keys-how-does-this-affect-industry-wide-software-supply-chain">
          <front>
            <title>Leaked MSI Source Code with Intel OEM Keys: How Does This Affect Industry-wide Software Supply Chain?</title>
            <author>
              <organization>Binarly efiXplorer Team</organization>
            </author>
            <date year="2025" month="May" day="27"/>
          </front>
        </reference>
        <reference anchor="CABFORUM" target="https://cabforum.org/wp-content/uploads/CA-Browser-Forum-BR-1.7.3.pdf">
          <front>
            <title>CA/Browser Forum Baseline Requirements for the Issuance and Management of Publicly-Trusted Certificates, v.1.7.3</title>
            <author>
              <organization>CA/Browser Forum</organization>
            </author>
            <date year="2020" month="October"/>
          </front>
        </reference>
        <reference anchor="TPMEK" target="https://trustedcomputinggroup.org/wp-content/uploads/TCG-EK-Credential-Profile-for-TPM-Family-2.0-Level-0-Version-2.6_pub.pdf">
          <front>
            <title>TCG Credential Profile EK 2.0</title>
            <author>
              <organization>Trusted Computing Group</organization>
            </author>
            <date year="2024" month="December" day="04"/>
          </front>
        </reference>
        <reference anchor="PUF" target="https://en.wikipedia.org/wiki/Physical_unclonable_function">
          <front>
            <title>Physically Uncloneable Function</title>
            <author>
              <organization>Wikipedia</organization>
            </author>
            <date year="2023" month="July" day="23"/>
          </front>
        </reference>
        <reference anchor="curl" target="https://curl.se/download.html">
          <front>
            <title>CURL</title>
            <author initials="D." surname="Stenberg" fullname="Daniel Stenberg">
              <organization/>
            </author>
            <date year="2025" month="May" day="28"/>
          </front>
        </reference>
        <reference anchor="measuredboot" target="https://trustedcomputinggroup.org/resource/pc-client-specific-platform-firmware-profile-specification/">
          <front>
            <title>TCG PC Client Specific Platform Firmware Profile Specification</title>
            <author>
              <organization>Trusted Computing Group</organization>
            </author>
            <date year="2023" month="December" day="04"/>
          </front>
        </reference>
        <reference anchor="TPM20spec" target="https://trustedcomputinggroup.org/resource/tpm-library-specification/">
          <front>
            <title>TPM 2.0 Library</title>
            <author>
              <organization>Trusted Computing Group</organization>
            </author>
            <date year="2025" month="March"/>
          </front>
        </reference>
        <reference anchor="ISO27001">
          <front>
            <title>ISO/IEC 27001:2022 Information security, cybersecurity and privacy protection -- Information security management systems -- Requirements</title>
            <author>
              <organization>International Organization for Standardization</organization>
            </author>
            <date year="2022"/>
          </front>
        </reference>
        <reference anchor="RFC9711">
          <front>
            <title>The Entity Attestation Token (EAT)</title>
            <author fullname="L. Lundblade" initials="L." surname="Lundblade"/>
            <author fullname="G. Mandyam" initials="G." surname="Mandyam"/>
            <author fullname="J. O'Donoghue" initials="J." surname="O'Donoghue"/>
            <author fullname="C. Wallace" initials="C." surname="Wallace"/>
            <date month="April" year="2025"/>
            <abstract>
              <t>An Entity Attestation Token (EAT) provides an attested claims set that describes the state and characteristics of an entity, a device such as a smartphone, an Internet of Things (IoT) device, network equipment, or such. This claims set is used by a relying party, server, or service to determine the type and degree of trust placed in the entity.</t>
              <t>An EAT is either a CBOR Web Token (CWT) or a JSON Web Token (JWT) with attestation-oriented claims.</t>
            </abstract>
          </front>
          <seriesInfo name="RFC" value="9711"/>
          <seriesInfo name="DOI" value="10.17487/RFC9711"/>
        </reference>
        <reference anchor="I-D.birkholz-suit-coswid-manifest">
          <front>
            <title>A SUIT Manifest Extension for Concise Software Identifiers</title>
            <author fullname="Henk Birkholz" initials="H." surname="Birkholz">
              <organization>Fraunhofer SIT</organization>
            </author>
            <date day="17" month="July" year="2018"/>
            <abstract>
              <t>   This document defines a resource extension for Concise Software
   Identifiers (CoSWID) that represents a SUIT firmware manifest.  This
   extension combines the information elements of the SUIT information
   model with the semantic expressiveness of Software Identifiers.  In
   consequence, this composite enables the integration of Firmware
   Updates for the Internet of Things (SUIT) in existing work-flows for
   updates of software components in general.

              </t>
            </abstract>
          </front>
          <seriesInfo name="Internet-Draft" value="draft-birkholz-suit-coswid-manifest-00"/>
        </reference>
        <reference anchor="I-D.ietf-iotops-mud-rats">
          <front>
            <title>MUD-Based RATS Resources Discovery</title>
            <author fullname="Henk Birkholz" initials="H." surname="Birkholz">
              <organization>Fraunhofer SIT</organization>
            </author>
            <author fullname="Michael Richardson" initials="M." surname="Richardson">
              <organization>Sandelman Software Works</organization>
            </author>
            <author fullname="Peter Chunchi Liu" initials="P. C." surname="Liu">
              <organization>Huawei Technologies</organization>
            </author>
            <date day="28" month="November" year="2025"/>
            <abstract>
              <t>   Manufacturer Usage Description (MUD) files and the MUD URIs that
   point to them are defined in RFC 8520.  This document introduces a
   new type of MUD file to be delivered in conjunction with a MUD file
   signature and/or to be referenced via a MUD URI embedded in other
   documents or messages, such as an IEEE 802.1AR Secure Device
   Identifier (DevID) or a CBOR Web Token (CWT).  These signed documents
   can be presented to other entities, e.g., a network management system
   or network path orchestrator.  If this entity also takes on the role
   of a verifier as defined by the IETF Remote ATtestation procedureS
   (RATS) architecture, this verifier can use the references included in
   the MUD file specified in this document to discover, for example,
   appropriate reference value providers, endorsement documents or
   endorsement distribution APIs, trust anchor stores, remote verifier
   services (sometimes referred to as Attestation Verification
   Services), or transparency logs.  All theses references in the MUD
   file pointing to resources and auxiliary RATS services can satisfy
   general RATS prerequisite by enabling discovery or improve discovery
   resilience of corresponding resources or services.

              </t>
            </abstract>
          </front>
          <seriesInfo name="Internet-Draft" value="draft-ietf-iotops-mud-rats-02"/>
        </reference>
        <reference anchor="RFC8520">
          <front>
            <title>Manufacturer Usage Description Specification</title>
            <author fullname="E. Lear" initials="E." surname="Lear"/>
            <author fullname="R. Droms" initials="R." surname="Droms"/>
            <author fullname="D. Romascanu" initials="D." surname="Romascanu"/>
            <date month="March" year="2019"/>
            <abstract>
              <t>This memo specifies a component-based architecture for Manufacturer Usage Descriptions (MUDs). The goal of MUD is to provide a means for end devices to signal to the network what sort of access and network functionality they require to properly function. The initial focus is on access control. Later work can delve into other aspects.</t>
              <t>This memo specifies two YANG modules, IPv4 and IPv6 DHCP options, a Link Layer Discovery Protocol (LLDP) TLV, a URL, an X.509 certificate extension, and a means to sign and verify the descriptions.</t>
            </abstract>
          </front>
          <seriesInfo name="RFC" value="8520"/>
          <seriesInfo name="DOI" value="10.17487/RFC8520"/>
        </reference>
        <reference anchor="RFC7168">
          <front>
            <title>The Hyper Text Coffee Pot Control Protocol for Tea Efflux Appliances (HTCPCP-TEA)</title>
            <author fullname="I. Nazar" initials="I." surname="Nazar"/>
            <date month="April" year="2014"/>
            <abstract>
              <t>The Hyper Text Coffee Pot Control Protocol (HTCPCP) specification does not allow for the brewing of tea, in all its variety and complexity. This paper outlines an extension to HTCPCP to allow for pots to provide networked tea-brewing facilities.</t>
            </abstract>
          </front>
          <seriesInfo name="RFC" value="7168"/>
          <seriesInfo name="DOI" value="10.17487/RFC7168"/>
        </reference>
        <reference anchor="RFC9641">
          <front>
            <title>A YANG Data Model for a Truststore</title>
            <author fullname="K. Watsen" initials="K." surname="Watsen"/>
            <date month="October" year="2024"/>
            <abstract>
              <t>This document presents a YANG module for configuring bags of certificates and bags of public keys that can be referenced by other data models for trust. Notifications are sent when certificates are about to expire.</t>
            </abstract>
          </front>
          <seriesInfo name="RFC" value="9641"/>
          <seriesInfo name="DOI" value="10.17487/RFC9641"/>
        </reference>
        <reference anchor="RFC9180">
          <front>
            <title>Hybrid Public Key Encryption</title>
            <author fullname="R. Barnes" initials="R." surname="Barnes"/>
            <author fullname="K. Bhargavan" initials="K." surname="Bhargavan"/>
            <author fullname="B. Lipp" initials="B." surname="Lipp"/>
            <author fullname="C. Wood" initials="C." surname="Wood"/>
            <date month="February" year="2022"/>
            <abstract>
              <t>This document describes a scheme for hybrid public key encryption (HPKE). This scheme provides a variant of public key encryption of arbitrary-sized plaintexts for a recipient public key. It also includes three authenticated variants, including one that authenticates possession of a pre-shared key and two optional ones that authenticate possession of a key encapsulation mechanism (KEM) private key. HPKE works for any combination of an asymmetric KEM, key derivation function (KDF), and authenticated encryption with additional data (AEAD) encryption function. Some authenticated variants may not be supported by all KEMs. We provide instantiations of the scheme using widely used and efficient primitives, such as Elliptic Curve Diffie-Hellman (ECDH) key agreement, HMAC-based key derivation function (HKDF), and SHA2.</t>
              <t>This document is a product of the Crypto Forum Research Group (CFRG) in the IRTF.</t>
            </abstract>
          </front>
          <seriesInfo name="RFC" value="9180"/>
          <seriesInfo name="DOI" value="10.17487/RFC9180"/>
        </reference>
        <reference anchor="RFC4086">
          <front>
            <title>Randomness Requirements for Security</title>
            <author fullname="D. Eastlake 3rd" initials="D." surname="Eastlake 3rd"/>
            <author fullname="J. Schiller" initials="J." surname="Schiller"/>
            <author fullname="S. Crocker" initials="S." surname="Crocker"/>
            <date month="June" year="2005"/>
            <abstract>
              <t>Security systems are built on strong cryptographic algorithms that foil pattern analysis attempts. However, the security of these systems is dependent on generating secret quantities for passwords, cryptographic keys, and similar quantities. The use of pseudo-random processes to generate secret quantities can result in pseudo-security. A sophisticated attacker may find it easier to reproduce the environment that produced the secret quantities and to search the resulting small set of possibilities than to locate the quantities in the whole of the potential number space.</t>
              <t>Choosing random quantities to foil a resourceful and motivated adversary is surprisingly difficult. This document points out many pitfalls in using poor entropy sources or traditional pseudo-random number generation techniques for generating such quantities. It recommends the use of truly random hardware techniques and shows that the existing hardware on many systems can be used for this purpose. It provides suggestions to ameliorate the problem when a hardware solution is not available, and it gives examples of how large such quantities need to be for some applications. This document specifies an Internet Best Current Practices for the Internet Community, and requests discussion and suggestions for improvements.</t>
            </abstract>
          </front>
          <seriesInfo name="BCP" value="106"/>
          <seriesInfo name="RFC" value="4086"/>
          <seriesInfo name="DOI" value="10.17487/RFC4086"/>
        </reference>
        <reference anchor="RFC9397">
          <front>
            <title>Trusted Execution Environment Provisioning (TEEP) Architecture</title>
            <author fullname="M. Pei" initials="M." surname="Pei"/>
            <author fullname="H. Tschofenig" initials="H." surname="Tschofenig"/>
            <author fullname="D. Thaler" initials="D." surname="Thaler"/>
            <author fullname="D. Wheeler" initials="D." surname="Wheeler"/>
            <date month="July" year="2023"/>
            <abstract>
              <t>A Trusted Execution Environment (TEE) is an environment that enforces the following: any code within the environment cannot be tampered with, and any data used by such code cannot be read or tampered with by any code outside the environment. This architecture document discusses the motivation for designing and standardizing a protocol for managing the lifecycle of Trusted Applications running inside such a TEE.</t>
            </abstract>
          </front>
          <seriesInfo name="RFC" value="9397"/>
          <seriesInfo name="DOI" value="10.17487/RFC9397"/>
        </reference>
      </references>
    </references>
  </back>
  <!-- ##markdown-source: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-->

</rfc>
