Merkle Tree Certificates

Merkle Tree Certificates Google LLC

davidben@google.com

Apple Inc.

asymmetric@apple.com

Cloudflare

bas@cloudflare.com

Cloudflare

lvalenta@cloudflare.com

Geomys

ietf@filippo.io

Security PKI, Logs, And Tree Signatures This document describes Merkle Tree certificates, a new form of X.509 certificates which integrate public logging of the certificate, in the style of Certificate Transparency. The integrated design reduces logging overhead in the face of both shorter-lived certificates and large post-quantum signature algorithms, while still achieving comparable security properties to existing X.509 constructions and Certificate Transparency. Merkle Tree certificates additionally admit an optional size optimization that avoids signatures altogether, at the cost of only applying to up-to-date relying parties and older certificates. About This Document The latest revision of this draft can be found at . Status information for this document may be found at . Discussion of this document takes place on the PKI, Logs, And Tree Signatures Working Group mailing list (), which is archived at . Subscribe at . Source for this draft and an issue tracker can be found at .

Introduction In Public Key Infrastructures (PKIs) that use Certificate Transparency (CT) for a public logging requirement, an authenticating party must present Signed Certificate Timestamps (SCTs) alongside certificates. CT policies often require two or more SCTs per certificate , each of which carries a signature. These signatures are in addition to those in the certificate chain itself. Current signature schemes can use as few as 32 bytes per key and 64 bytes per signature , but post-quantum replacements are much larger. For example, ML-DSA-44 uses 1,312 bytes per public key and 2,420 bytes per signature. ML-DSA-65 uses 1,952 bytes per public key and 3,309 bytes per signature. Even with a directly-trusted intermediate (), two SCTs and a leaf certificate signature add 7,260 bytes of authentication overhead with ML-DSA-44 and 9,927 bytes with ML-DSA-65. This increased overhead additionally impacts CT logs themselves. Most of a log's costs scale with the total storage size of the log. Each log entry contains both a public key, and a signature from the CA. With larger public keys and signatures, the size of each log entry will grow. Additionally, as PKIs transition to shorter-lived certificates , the rate at which entries are added to the log will increase. This document introduces Merkle Tree Certificates (MTCs), a new form of X.509 certificate that integrates logging with certificate issuance. Each CA maintains logs of everything it issues, signing views of its logs to assert it has issued the contents. The CA signature is combined with cosignatures from other parties who verify correct operation and optionally mirror the logs. These signatures, together with an inclusion proof for an individual entry, constitute a certificate. This achieves the following:

Log entries do not scale with public key and signature sizes. Entries replace public keys with hashes and do not contain signatures, while preserving non-repudiability ().
Long-expired entries can be revoked from relying parties in bulk (see ). This allows logging and monitoring infrastructure to scale by retention policies, not the lifetime of the log, even as certificate lifetimes decrease.
After a processing delay, authenticating parties can obtain a second "landmark-relative" certificate for the same log entry. This second certificate is an optional size optimization that avoids the need for any signatures, assuming an up-to-date client that has some predistributed log information.

gives an overview of the system. describes a Merkle Tree primitive used by this system. describes the log structure. Finally, and describe how to construct and consume a Merkle Tree certificate.

Conventions and Definitions The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "NOT RECOMMENDED", "MAY", and "OPTIONAL" in this document are to be interpreted as described in BCP 14 when, and only when, they appear in all capitals, as shown here. This document additionally uses the TLS presentation language defined in , as well as the notation defined in . It extends the numeric types defined in with a big-endian, 48-bit integer: U+ followed by four hexadecimal characters denotes a Unicode codepoint, to be encoded in UTF-8 . 0x followed by two hexadecimal characters denotes a byte value in the 0-255 range. The decimal representation of a non-negative integer is its base-ten representation, written with the ASCII digits 0 through 9 (U+0030 through U+0039). Zero is written as the single digit 0, and no other value is written with a leading 0. [start, end), where start <= end, denotes the half-open interval containing integers x such that start <= x < end. Given a non-negative integer n,

LSB(n) refers to the least-significant bit of n's binary representation. Equivalently, it is the remainder when n is divided by 2.
BIT_WIDTH(n) refers to the smallest number of bits needed to represent n. BIT_WIDTH(0) is zero.
POPCOUNT(n) refers to the number of set bits in n's binary representation.
BIT_CEIL(n) refers to the smallest power of 2 that is greater or equal to n.

To left-shift a non-negative integer n is to shift each bit in its binary representation to one upper position. Equivalently, it is n times 2. Given non-negative integers a and b, a << b refers to a left-shifted b times. To right-shift a non-negative integer n is to shift each bit in its binary representation to one lower position, discarding the least-significant bit. Equivalently, it is the floor of n divided by 2. Given non-negative integers a and b, a >> b refers to a right-shifted b times. Given two non-negative integers a and b, a & b refers to the non-negative integer such that each bit position is set if the corresponding bit is set in both a and b, and unset otherwise. This is commonly referred to as the bitwise AND operator.

Terminology and Roles This document discusses the following roles:

Authenticating party:: The party that authenticates itself in the protocol. In TLS, this is the side sending the Certificate and CertificateVerify message.
Certification authority (CA):: The service that issues certificates to the authenticating party, after performing some validation process on the certificate contents.
Relying party:: The party to whom the authenticating party presents its identity. In TLS, this is the side receiving the Certificate and CertificateVerify message.
Monitor:: Parties who watch logs for certificates of interest, analogous to the role in .
Issuance log:: A log, maintained by the CA, containing certification statements issued by that CA. A CA operates some number of issuance logs, which together contain all statements issued by that CA.
Cosigner:: A service that signs views of an issuance log, to assert correct operation and other properties about the entries.

Additionally, there are several terms used throughout this document to describe this proposal. This section provides an overview. They will be further defined and discussed in detail throughout the document.

Checkpoint:: A description of the complete state of the log at some time.
Entry:: An individual element of the log, describing information which the CA has validated and certified.
Subtree:: A smaller Merkle Tree over a portion of the log, defined by an interior node of some snapshot of the log. Subtrees can be efficiently shown to be consistent with the whole log.
Inclusion proof:: A sequence of hashes that efficiently proves some entry is contained in some checkpoint or subtree.
Consistency proof:: A sequence of hashes that efficiently proves a checkpoint or subtree is contained within another checkpoint.
Cosignature:: A signature from either the CA or other cosigner, over some checkpoint or subtree.
Landmark:: One of a sequence of tree sizes, infrequently chosen and used to calculate landmark subtrees for predistribution to relying parties.
Landmark subtree:: One of the two (possibly empty) subtrees determined by an interval between two landmarks. Predistributed and used as the basis for landmark-relative certificates.
Standalone certificate:: A certificate containing an inclusion proof to some subtree, and several cosignatures over that subtree.
Landmark-relative certificate:: An optimized certificate containing an inclusion proof to a landmark subtree, and no signatures.
Directly-signed certificate:: A certificate issued using the existing, non-MTC construction, where the TBSCertificate is passed directly to the private key's signing operation.

Overview In Certificate Transparency, a CA first certifies information by signing it, then submits the resulting certificate (or precertificate) to logs for logging. Merkle Tree Certificates invert this process: the CA certifies information by logging it, then submits the log to cosigners to verify log operation. A certificate is assembled from the result and proves the information is in the CA's log.

A diagram of the MTC issuance architecture, detailed below 5. Download certificates | | / \ | | | | / \ | | * tbscert | | +-------+ | | = = = inclusion proof | | * * * * tbscert entries | | [ CA ] cosignatures | | | | [ mirror ] | +------------------------------+ +-----------------------------+ / | \ / | \ 4. Submit log to cosigners V V V for cosignatures +-- Mirrors, other cosigners --+ +-- Monitors -----------------+ | | | | | +---[ CA cosign ] +-+ | | | / \ [ mirror cosign ] | | | | | / \ | | | | | / \ <-+-+--+-- 6. Monitor CA operation | | +-------+ | | | | | * * * * | | +-----------------------------+ +-+----------------------------+ | | ...quorum of cosigners... | +------------------------------+ ]]> Merkle Tree Certificates are issued as follows. depicts this process.

The authenticating party requests a certificate, e.g. over ACME
The CA validates each incoming issuance request, e.g. with ACME challenges. From there, the process diverges from CT-based PKIs.
The CA operates a series of append-only issuance logs (). Unlike a CT log, these logs only contain entries added by the CA:
1. The CA adds a TBSCertificateLogEntry (, abbreviated "tbscert entries" in the diagram) to an issuance log, describing the information it is certifying.
2. The CA signs a checkpoint, which describes the current state of the log. A signed checkpoint certifies that the CA issued every entry in the Merkle Tree ().
3. The CA additionally signs subtrees () that together contain certificates added since the last checkpoint (). This is an optimization to reduce inclusion proof sizes. A signed subtree certifies that the CA has issued every entry in the subtree.
The CA submits the new log state to cosigners. Cosigners validate the log is append-only and optionally provide additional services, such as mirroring its contents. They cosign the CA's checkpoints and subtrees.
The CA now has enough information to construct a certificate and give it to the authenticating party. A certificate contains:
- The TBSCertificate being certified
- An inclusion proof from the TBSCertificate to some subtree
- Cosignatures from the CA and cosigners on the subtree
As in Certificate Transparency, monitors observe the CA's issuance logs to ensure the CA is operated correctly.

A certificate with cosignatures is known as a standalone certificate. Analogous to X.509 trust anchors and trusted CT logs, relying parties are configured with trusted cosigners () that allow them to accept Merkle Tree certificates. The inclusion proof proves the TBSCertificate is part of some subtree, and cosignatures from trusted cosigners prove the subtree was certified by the CA and available to monitors. Where CT logs entire certificates, the issuance log's entries are smaller TBSCertificateLogEntry () structures, which do not scale with public key or signature size. This same issuance process also produces a landmark-relative certificate. This is an optional, optimized certificate that avoids all cosignatures, including the CA signature. Landmark-relative certificates are available after a short period of time and usable with up-to-date relying parties.

Periodically, the tree size of the CA's most recent checkpoint is designated as a landmark. This determines landmark subtrees, which are common points of reference between relying parties and landmark-relative certificates.
Once some landmark includes the TBSCertificate, the landmark-relative certificate is constructed with:
- The TBSCertificate being certified
- An inclusion proof from the TBSCertificate to a landmark subtree
In the background, landmark subtrees are predistributed to relying parties, with cosignatures checked against relying party requirements. This occurs periodically in the background, separate from the application protocol.
During the application protocol, such as TLS , if the relying party already supports the landmark subtree, the authenticating party can present the landmark-relative certificate. Otherwise, it presents a standalone certificate. The authenticating party may also select between several landmark-relative certificates, as described in .

Subtrees This section extends the Merkle Tree definition in by defining a subtree of a Merkle Tree. A subtree is itself a Merkle Tree, built over an interval of entries from the original tree. defines a subtree formally, including the constraints on those intervals. As with Merkle Trees, a subtree inclusion proof, defined in , can prove an entry is contained in some subtree. Subtrees, and thus their inclusion proofs, are smaller than those of the original tree, so this document uses subtree inclusion proofs as a certificate size optimization. Not all intervals can form subtrees. Subtrees are limited to intervals that can be efficiently proven consistent with the original tree, using subtree consistency proofs defined in . However, every interval of a Merkle Tree can be efficiently covered by two subtrees. describes how to determine these subtrees. and provide test vectors for the algorithms defined in this section.

Definition of a Subtree Given an ordered list of n inputs, D_n = {d[0], d[1], ..., d[n-1]}, defines the Merkle Tree via the Merkle Tree Hash MTH(D_n). A subtree of this Merkle Tree is itself a Merkle Tree, defined by MTH(D[start:end]). start and end are integers such that:

0 <= start <= end <= n
start is a multiple of BIT_CEIL(end - start)

The second condition ensures that MTH(D[start:end]), built over D[start:end] as an independent list, is sufficiently aligned with the original Merkle Tree to support subtree consistency proofs. See for more details. In implementations using fixed-width integers, BIT_CEIL(end - start) above may exceed end and potentially overflow. For example, if start is zero and end is 2⁶³+1, BIT_CEIL(end - start) is 2⁶⁴. The following is an example C++ implementation that handles this condition. end) { return false; } uint64_t size = end - start; if (size > (uint64_t{1} << 63)) { return start == 0; // bit_ceil below will overflow. } return (start & (std::bit_ceil(size) - 1)) == 0; } ]]> For all x, [0, x) is a valid subtree (0 is a multiple of everything), and [x, x) is a valid subtree (BIT_CEIL(0) is 1). The size of the subtree is end - start. In the context of a single Merkle Tree, this document denotes subtree MTH(D[start:end]) by half-open interval [start, end). It contains the entries whose indices are in that half-open interval. As a Merkle Tree grows, its subtrees remain unchanged. That is, if end <= m <= n, the subtree [start, end) of MTH(D[0:m]) and the subtree [start, end) of MTH(D_n) are both valid and identical.

Example Subtrees shows the subtrees [4, 8) and [8, 13):

Two example subtrees Both can be viewed as subtrees of a Merkle Tree of size 13, depicted in . Nodes in common with [4, 8) and [8, 13) are marked with doubled and wavy lines, respectively.

A Merkle Tree of size 13 In some cases, not every node of a subtree will appear in the larger Merkle Tree. depicts a Merkle Tree of size 14. Nodes in common with [4, 8) and [8, 13) are marked as above. While all nodes of [4, 8) appear in the tree, non-leaf nodes on [8, 13)'s right edge do not. However, there is still sufficient overlap to construct subtree consistency proofs ().

A Merkle Tree of size 14 discusses subtrees in more detail.

Subtree Inclusion Proofs Subtrees are Merkle Trees, so entries can be proven to be contained in the subtree. A subtree inclusion proof for entry index of the subtree [start, end) is a Merkle inclusion proof, as defined in , where m is index - start and the tree inputs are D[start:end]. Subtree inclusion proofs contain a sequence of nodes that are sufficient to reconstruct the subtree hash, MTH(D[start:end]), out of the hash for entry index, MTH({d[index]}), thus demonstrating that the subtree hash contains the entry's hash. A subtree inclusion proof for a subtree of size n contains at most ceil(log2(n)) hashes, or BIT_WIDTH(n - 1) hashes.

Example Subtree Inclusion Proofs The inclusion proof for entry 10 of subtree [8, 13) contains the hashes MTH({d[11]}), MTH(D[8:10]), and MTH({d[12]}), depicted in . MTH({d[10]}) is not part of the proof because the verifier is assumed to already know its value.

An example subtree inclusion proof

Evaluating a Subtree Inclusion Proof Given a subtree inclusion proof, inclusion_proof, for entry index, with hash entry_hash, of a subtree [start, end), the subtree inclusion proof can be evaluated to compute the expected subtree hash:

Check that [start, end) is a valid subtree (), and that start <= index < end. If either does not hold, fail proof evaluation.
Set fn to index - start and sn to end - start - 1.
Set r to entry_hash.
For each value p in the inclusion_proof array:
1. If sn is 0, then stop the iteration and fail proof evaluation.
2. If LSB(fn) is set, or if fn is equal to sn, then:
  1. Set r to HASH(0x01 || p || r).
  2. Until LSB(fn) is set, right-shift fn and sn equally.
  Otherwise:
  1. Set r to HASH(0x01 || r || p).
3. Finally, right-shift both fn and sn one time.
If sn is not zero, fail proof evaluation.
Return r as the expected subtree hash.

This is the same as the procedure in , where leaf_index is index - start, tree_size is end - start, and r is returned instead of compared with root_hash. explains this procedure in more detail.

Verifying a Subtree Inclusion Proof Given a subtree inclusion proof, inclusion_proof, for entry index, with hash entry_hash, of a subtree [start, end) with hash subtree_hash, the subtree inclusion proof can be verified to verify the described entry is contained in the subtree:

Let expected_subtree_hash be the result of evaluating the inclusion proof as described . If evaluation fails, fail the proof verification.
If subtree_hash is equal to expected_subtree_hash, the entry is contained in the subtree. Otherwise, fail the proof verification.

Subtree Consistency Proofs A subtree [start, end) can be efficiently proven to be consistent with the full Merkle Tree. That is, given MTH(D[start:end]) and MTH(D_n), the proof demonstrates that the input D[start:end] to the subtree hash was equal to the corresponding elements of the input D_n to the Merkle Tree hash. Subtree consistency proofs contain sufficient nodes to reconstruct both the subtree hash, MTH(D[start:end]), and the original tree hash, MTH(D_n), in such a way that every input to the subtree hash was also incorporated into the original tree hash.

Generating a Subtree Consistency Proof The subtree consistency proof, SUBTREE_PROOF(start, end, D_n) is defined similarly to . If start = end, the consistency proof is empty: Otherwise, start < end and SUBTREE_PROOF is defined by a helper function: The boolean parameter tracks whether the first hash in the proof will be omitted by the base case. The first hash is omitted when it's equal to the original subtree hash MTH(D[start:end]), since the verifier will already know that hash. That happens when the original subtree's root is a node in the Merkle Tree constructed from D_n, or equivalently, when the original subtree is full or has end = n. If start = 0 and end = n, the subtree is the root (base case): Otherwise, n > 1. Let k be the largest power of two smaller than n. The consistency proof is defined recursively as:

If end <= k, the subtree is on the left of k. The proof proves consistency with the left child and includes the right child:
If k <= start, the subtree is on the right of k. The proof proves consistency with the right child and includes the left child.
Otherwise, start < k < end, which implies start = 0. The proof proves consistency with the right child and includes the left child.

When start is zero, this computes a Merkle consistency proof: When end = start + 1, this computes a Merkle inclusion proof: explains the structure of a subtree consistency proof in more detail.

Example Subtree Consistency Proofs The subtree consistency proof for [4, 8) and a tree of size 14 contains MTH(D[0:4]) and MTH(D[8:14]), depicted in with doubled lines. The verifier is assumed to know the subtree hash, so there is no need to include MTH(D[4:8]), depicted with wavy lines, in the consistency proof.

An example subtree consistency proof that begins at the root of the subtree The subtree consistency proof for [8, 13) and a tree of size 14 contains MTH({d[12]}), MTH({d[13]}), MTH(D[8:12]), and MTH(D[0:8]), depicted in with doubled lines. Not every node in [8, 13) is also in the overall tree, so the proof must include sufficient nodes to reconstruct both hashes. However, there is enough overlap for the proof to be possible.

An example subtree consistency proof that decomposes the subtree

Verifying a Subtree Consistency Proof The following procedure can be used to verify a subtree consistency proof. Given a Merkle Tree over n elements, a subtree defined by [start, end), a consistency proof proof, a subtree hash node_hash, and a root hash root_hash:

Check that [start, end) is a valid subtree (), and that end <= n. If either does not hold, fail proof verification. These checks imply 0 <= start <= end <= n.
If start equals end, check the following conditions:
- proof is an empty array.
- node_hash is equal to HASH(), the hash of the empty string.
If either condition does not hold, stop and fail the proof verification. If both hold, stop and accept the proof.
Set fn to start, sn to end - 1, and tn to n - 1.
If sn is tn, then:
1. Until fn is sn, right-shift fn, sn, and tn equally.
Otherwise:
1. Until fn is sn or LSB(sn) is not set, right-shift fn, sn, and tn equally.
If fn is sn, set fr and sr to node_hash.
Otherwise:
1. If proof is an empty array, stop and fail verification.
2. Remove the first value of the proof array and set fr and sr to the removed value.
For each value c in the proof array:
1. If tn is 0, then stop the iteration and fail the proof verification.
2. If LSB(sn) is set, or if sn is equal to tn, then:
  1. If fn < sn, set fr to HASH(0x01 || c || fr).
  2. Set sr to HASH(0x01 || c || sr).
  3. Until LSB(sn) is set, right-shift fn, sn, and tn equally.
3. Otherwise:
  1. Set sr to HASH(0x01 || sr || c).
4. Right-shift fn, sn, and tn once more.
Compare tn to 0, fr to node_hash, and sr to root_hash. If any are not equal, fail the proof verification. If all are equal, accept the proof.

explains this procedure in more detail.

Efficiently Covering Arbitrary Intervals This document uses subtrees to sign over arbitrary intervals, [start, end), of a Merkle Tree. However, not all intervals are valid subtrees. While a protocol could build a smaller Merkle Tree, MTH(D[start:end]), and compute inclusion proofs of any element, this smaller Merkle Tree cannot, in general, be efficiently proven consistent with the overall Merkle Tree. Enlarging the interval to a valid subtree would mitigate this. However, the smallest subtree containing [start, end) may be much larger than [start, end). For example, shows the smallest subtree that contains [7, 9) in a 9-element tree. The smallest single subtree that contains the interval is [0, 9), but this is the entire tree.

An example showing an inefficient choice of a single subtree While one subtree can be inefficient, two subtrees are sufficient to efficiently cover any interval, as described below.

Selecting Two Subtrees Given any interval, [start, end), this section defines a procedure for selecting two subtrees, left and right, such that:

left and right are valid subtrees, so it is possible to compute subtree consistency proofs.
The disjoint union of left, followed by right, contains [start, end). That is, left.start <= start <= left.end = right.start <= end <= right.end.
While left may contain extra elements before start, right does not contain any extra elements. That is, end = right.end.
Each subtree's size is at most BIT_CEIL(end - start).

The pair of subtree hashes for left and right can support inclusion proofs for any element of [start, end). The largest such inclusion proof is no bigger than the largest inclusion proof in MTH(D[start:end]). Unlike MTH(D[start:end]), these subtree hashes can be shown consistent with the overall Merkle Tree using subtree consistency proofs. The subtrees are selected as follows:

If end - start is less than or equal to 1, return the subtrees [start, end) and [end, end).
Otherwise:
1. Let last be end - 1, the last index in [start, end).
2. Let split be the bit index of the most significant bit where start and last differ. Bits are numbered from the least significant bit, starting at zero. split is the height at which start and last's paths in the tree diverge.
3. Let mid be last with the least significant split bits set to zero. mid is the leftmost leaf node in the above divergence point's right branch.
4. Within the least significant split bits of start, let b be the bit index of the most significant bit with value zero, if any:
  1. If there is such a bit, let left_split be b + 1.
  2. Otherwise, let left_split be zero.
  left_split is the height of the lowest common ancestor of the nodes in [start, mid).
5. Let left_start be start with the least significant left_split bits set to zero. left_start is the above lowest common ancestor's leftmost leaf node.
6. Return the subtrees [left_start, mid) and [mid, end).

Intuitively, this procedure considers the tree MTH(D[0:end]) and finds the lowest common ancestor of the elements in [start, end). It splits the interval by that ancestor's left and right children and returns the lowest common ancestor of each half. The following Python code implements this procedure: shows the subtrees which cover [5, 13) in a Merkle Tree of 13 elements in wavy lines. The two subtrees selected are [4, 8) and [8, 13). Note that the subtrees cover a slightly larger interval than [5, 13).

An example selection of subtrees to cover an interval

Certification Authorities A CA consists of the following components:

A CA ID (), which uniquely identifies the CA.
A collision-resistant cryptographic hash function, used by the CA's issuance logs. SHA-256 is RECOMMENDED. Throughout this document, this hash function is referred to as HASH, and the size of its output in bytes is referred to as HASH_SIZE.
A series of issuance logs (), which contain all statements the CA has certified. One issuance log is designated as the current log.
A CA cosigner (), which signs subtrees of issuance logs to certify their contents.
Optionally, a landmark sequence per log (), to support optimized landmark-relative certificates.

defines an X.509 certificate representation of a CA.

Certification Authority Identifiers Each Merkle Tree Certificate CA has a CA ID to identify it. This CA ID is a trust anchor ID . Once allocated, the ID's entire object identifier (OID) arc is reserved by this protocol. Given a CA ID whose OID representation is caID, this document allocates the following OIDs:

For each positive integer N, the OID {caID logs(0) N} represents the issuance log N ().
For each positive integer N and L, the OID {caID landmarks(1) N L} represents landmark L () of issuance log N. These OIDs may be used as trust anchor IDs, as described in . These OIDs are used when it is necessary to identify an individual landmark, e.g. as in the recovery mechanism described in .
For each positive integer N and L, the OID {caID landmarkGroups(2) N L} represents a trust anchor group () containing landmark L of log N and earlier landmarks of that log, as defined in . These OIDs may be used to advertise a series of landmarks at once.

Future extensions to this protocol MAY define further allocations. A CA ID determines a PKIX distinguished name () that can be used in the issuer or subject field of an X.509 TBSCertificate. This distinguished name has a single relative distinguished name, which has a single attribute. The attribute has type id-rdna-trustAnchorID, defined below: The attribute's value is a RELATIVE-OID containing the trust anchor ID's ASN.1 representation. For example, the distinguished name for a CA with ID 32473.1 would be represented in syntax of as: For initial experimentation, early implementations of this design will use the OID 1.3.6.1.4.1.44363.47.3 instead of id-rdna-trustAnchorID. Cloudflare has kindly donated the 1.3.6.1.4.1.44363.47 OID arc for use in this document. For example, the distinguished name for an experimental CA with ID 32473.1 would be represented in syntax of as:

Issuance Logs A CA operates a series of issuance logs, each identified by a positive integer log number. Log numbers are numbered consecutively from 1 to at most 65535 (2¹⁶-1). Each issuance log has a log ID, which is a trust anchor ID constructed by concatenating the following OID components:

The CA ID ()
The constant 0
The log number of the log

A log ID specifies both the CA and the log number in a single ID. Each issuance log describes an append-only sequence of at most 2⁴⁸-1 entries (). Each entry is identified by an integer index, assigned consecutively starting from zero. Each entry is an assertion that the CA has certified. The entries in the issuance log are represented as a Merkle Tree, described in . Unlike and , an issuance log does not have a public submission interface. The log only contains entries which the log operator, i.e. the CA, chose to add. As entries are added, the Merkle Tree is updated to be computed over the new sequence. A snapshot of the log is known as a checkpoint. A checkpoint is identified by its tree size, that is the number of elements committed to the log at the time. Its contents can be described by the Merkle Tree Hash () of entries zero through tree_size - 1. At any point in time, one of the CA's issuance logs is its current log. Initially, this is log 1. A CA MUST NOT append to any log that is not the current log. Logs before the current log may have historical entries. Logs after the current log MUST be empty. A CA MAY increment its current log number as part of recovering from certain operational failures. See for further discussion.

Log Entries Each entry in the log is an MTCLogEntry, defined with the TLS presentation syntax below. An MTCLogEntry describes certificate information that the CA has validated and certified. ; } MTCLogEntryExtension; enum { null_entry(0), tbs_cert_entry(1), (2^16-1) } MTCLogEntryType; struct { MTCLogEntryExtension extensions<0..2^16-1>; MTCLogEntryType type; select (type) { case null_entry: Empty; case tbs_cert_entry: opaque tbs_cert_entry_data[N]; /* May be extended with future types. */ } } MTCLogEntry; ]]> The extensions field is a list of tag-length-value extensions associated with the log entry. Extensions MUST appear in the list in ascending order by extension_type, and the list MUST NOT contain two extensions with the same extension_type. When type is null_entry, the entry does not represent any information. Entries at any index in the log MAY have type null_entry. When type is tbs_cert_entry, N is the number of bytes needed to consume the rest of the input. An MTCLogEntry is expected to be decoded in contexts where the total length of the entry is known. tbs_cert_entry_data contains the contents octets (i.e. excluding the initial identifier and length octets) of the DER encoding of a TBSCertificateLogEntry, defined below. Equivalently, tbs_cert_entry_data contains the DER encodings of each field of the TBSCertificateLogEntry, concatenated. This construction allows a single-pass implementation in . The fields of a TBSCertificateLogEntry are defined as follows:

version, validity, subject, issuerUniqueID, subjectUniqueID, and extensions have the same semantics as the corresponding TBSCertificate fields, defined in .
issuer is the CA ID as a PKIX distinguished name, as described in .
- The issuer field is not human-readable. A TBSCertificateLogEntry MAY carry a human-readable label for the CA, suitable for display in user interfaces, in an issuer alternative name extension (). If present, the extension MUST be marked non-critical. The IssuerAltName SEQUENCE MUST contain a single GeneralName of type directoryName, whose Name MUST use the rdnSequence CHOICE. Each RelativeDistinguishedName MUST contain a single AttributeTypeAndValue. The extension is purely cosmetic, and MUST NOT be used in path validation or any other trust decision. The value MUST NOT be assumed unique across issuance logs and MAY change across entries in the same issuance log.
subjectPublicKeyAlgorithm describes the algorithm of the subject's public key. It is constructed identically to the algorithm field of a SubjectPublicKeyInfo ().
subjectPublicKeyInfoHash contains the hash of the subject's public key, encoded as a SubjectPublicKeyInfo. The hash uses the CA's hash function () and is computed over the SubjectPublicKeyInfo's DER encoding.

Note the subject's public key algorithm is incorporated into both subjectPublicKeyAlgorithm and subjectPublicKeyInfoHash. MTCLogEntry is an extensible structure. Future documents MAY define new values for MTCLogEntryType or MTCLogEntryExtensionType, with corresponding semantics. See and for additional discussion. An MTCLogEntry's size MUST NOT exceed 65535 (2¹⁶-1) bytes. TBSCertificateLogEntry does not include signatures and hashes public keys, so post-quantum algorithms do not contribute to this size.

Publishing Logs This protocol aims to enable monitors to detect misissued certificates by observing the issuance log. See . This document does not prescribe a particular method of observing the issuance log. The access protocols do not affect certificate interoperability, and different applications could have different needs. For example, a PKI that authenticates public services might publicly serve issuance logs, while a PKI that authenticates a single organization's intranet services might keep the log private to the organization. Relying parties SHOULD define log serving requirements, including the allowed protocols and expected availability, as part of their policies on which CAs to support. See also . If a serving protocol supports serving only a portion of the log, relying party policies SHOULD include requirements on which portions to serve. For example, defines a profile for Merkle Tree Certificates that uses .

Cosigners This section defines a log cosigner. A cosigner follows some append-only view of the log and signs subtrees () consistent with that view. The signatures generated by a cosigner are known as cosignatures. All subtrees signed by a cosigner MUST be consistent with each other. The cosigner may be external to the log, in which case it might ensure consistency by checking consistency proofs. The cosigner may be operated together with the log, in which case it can trust its log state. A cosignature MAY implicitly make additional statements about a subtree, determined by the cosigner's role. This document defines one concrete cosigner role, a CA cosigner (), to authenticate the log and certify entries. Other documents and specific deployments MAY define other cosigner roles, to perform different functions in a PKI. For example, defines a cosigner that only checks the log is append-only, and defines a cosigner that mirrors a log. Each cosigner has a public key and a cosigner ID, which uniquely identifies the cosigner. The cosigner ID is a trust anchor ID . By identifying the cosigner, the cosigner ID specifies the public key, signature algorithm, and any additional statements made by the cosigner's signatures. If a single operator performs multiple cosigner roles in an ecosystem, each role MUST use a distinct cosigner ID and SHOULD use a distinct key. Following the principle of key separation , cosigner keys SHOULD NOT be used for purposes outside this document. Additional uses MAY be defined but MUST NOT overlap with the signature format defined in . See for additional discussion. A single cosigner, with a single cosigner ID and public key, MAY generate cosignatures for multiple logs. In this case, signed subtrees only need to be consistent with others for the same log.

Signature Format A cosigner computes a subtree signature for a subtree in a log by signing a CosignedMessage, defined below using the TLS presentation language (): ; uint64 timestamp; opaque log_origin<1..2^8-1>; uint64 start; uint64 end; HashValue subtree_hash; } CosignedMessage; ]]> This signature format is designed to be compatible with the ML-DSA-44 signature construction in , but it supports signature algorithms other than ML-DSA-44 and tree hashes other than SHA-256. label is a fixed prefix for domain separation. Its value MUST be the string subtree/v1, followed by a newline (U+000A), followed by a zero byte (U+0000). cosigner_name and log_origin are computed from the cosigner ID and the issuance log's ID (), respectively. They contain the concatenation of:

The 16-byte ASCII string oid/1.3.6.1.4.1.
The trust anchor ID's ASCII representation ()

This is equivalent to the concatenation of:

The four-byte ASCII string oid/
The trust anchor ID as a full OID, in dotted decimal notation

For example, the trust anchor ID 32473.1 would be encoded as the ASCII string oid/1.3.6.1.4.1.32473.1. start and end MUST define a valid subtree of the log, and subtree_hash MUST be the subtree's hash value in the cosigner's view of the log. See . If timestamp is non-zero, it MUST be the time that the signature was produced. This time is represented as seconds since the Epoch, as defined in Section 4.19 of Volume 1 of . Additionally, if timestamp is non-zero, the following MUST be true:

start MUST be zero.
end MUST be the size of the largest consistent tree that the cosigner has observed for the log.

timestamp MAY be zero, in which case no additional constraints are placed on start or end (beyond being a valid subtree), and no statement is made about the signing time or largest observed tree.

Signature Semantics Before signing a subtree of some log, the cosigner MUST ensure that subtree_hash is consistent with its view of the log. Different cosigner roles will obtain this assurance differently. For example:

A cosigner MAY maintain a full copy of the log, e.g. if it's the log operator. The cosigner can then compute subtree_hash from this copy.
A cosigner MAY maintain the hash of the largest consistent tree observed by the log. The cosigner can then check subtree_hash with a subtree consistency proof ().
A cosigner MAY maintain any other representation of the log which allows it to verify the consistency of the log.

In all cases, the cosigner MUST ensure that, as it updates its view of the log, the old and new views are consistent. When a cosigner signs a subtree, it is held separately responsible both for the subtree being consistent with its other signatures, and for the cosigner-specific additional statements. That is, if a cosigner signs an inconsistent subtree, it is held responsible for its additional statements on all entries in the inconsistent subtree, even if some other signed subtree exists that asserts different entries. Subtree signatures can be used to sign timestamped log checkpoints by using a non-zero timestamp. A signature with a non-zero timestamp asserts the complete state of the cosigner's view of the log at a given time. These signatures are not directly used in Merkle Tree Certificates (), but cosigners MAY generate them, subject to the rules above, as part of other functions in a PKI, such as log serving or integrating an issuance log into a transparency ecosystem. For example, and use such signatures.

Signature Algorithms The cosigner's public key specifies both the key material and the signature algorithm to use with the key material. In order to change key or signature parameters, a cosigner operator MUST deploy a new cosigner, with a new cosigner ID. Signature algorithms MUST fully specify the algorithm parameters, such as hash functions used. Signatures are computed over the CosignedMessage described in . Log clients that accept cosignatures from some cosigner are assumed to be configured with all parameters necessary to verify that cosigner's signatures, including the signature algorithm and version of the signature format.

Certification Authority Cosigners A CA cosigner is a cosigner () that certifies the contents of a log. Each CA MUST operate a CA cosigner whose cosigner ID is the same as its CA ID (). A CA cosigner MUST NOT sign checkpoints or subtrees for logs not part of this CA instance. When a CA cosigner signs a subtree, it makes the additional statement that it has certified each entry in the subtree. For example, a domain-validating CA states that it has performed domain validation for each entry, at some time consistent with the entry's validity dates. CAs are held responsible for every entry in every subtree they sign. Proving an entry is included () in a CA-signed subtree is sufficient to prove the CA certified it. What it means to certify an entry depends on the entry type:

To certify an entry of type null_entry is a no-op. A CA MAY freely certify null_entry without being held responsible for any validation.
To certify an entry of type tbs_cert_entry is to certify the TBSCertificateLogEntry, as defined in .

Entries are extensible. Future documents MAY define type and extension_type values and the semantics of the data that they contain. A CA MUST NOT sign a subtree if it contains an entry with type or extension_type that it does not recognize. Doing so would certify that the CA has validated the information in some not-yet-defined format. further discusses security implications of such extensions. If the CA issues certificate revocation lists (CRLs) or Online Certificate Status Protocol (OCSP) responses , the CA's cosigner key MAY be used to directly sign TBSCertList or OCSP ResponseData structures, respectively, but only for this CA instance. Such uses remain subject to other X.509 constraints, such as the key usage extension, which are out of scope for this document. See for a discussion of domain separation. If the CA operator additionally operates a directly-signing X.509 CA, that CA key MUST be distinct from any Merkle Tree CA cosigner keys. In particular, a CA cosigner key MUST NOT be used to directly sign TBSCertificate structures. A CA cosigner key issues certificates by signing subtrees.

Representing Certification Authorities This section defines the X.509 Certificate representation of a Merkle Tree CA. It identifies the CA cosigner () and associated issuance logs. This information is encoded as follows:

The subject field MUST be the CA ID as a PKIX distinguished name, as described in .
The subjectPublicKeyInfo field MUST be the public key of the CA cosigner .
The extensions field MUST contain a critical Merkle Tree CA extension. This is defined below.
The subject key identifier extension (), if present, SHOULD be set to the CA ID . The CA ID is encoded in its binary representation, as defined in .

Other fields and extensions in apply unmodified. In particular:

The key usage extension () MUST be present and assert at least the keyCertSign bit.
The basic constraints extension () MUST be present and set the cA field to TRUE.

The Merkle Tree CA extension defines the remaining parameters specific to this protocol. It indicates that the subject of the certificate is a CA that issues Merkle Tree Certificates. If present, it MUST be marked as critical. The extension type identifies the Merkle Tree construction, and the contents define additional parameters of the CA cosigner. This document defines one extension type, id-pe-mtcCertificationAuthority-SHA256, which indicates hashing with SHA-256 . Other documents MAY define corresponding extensions for other hash functions or new versions of the tree construction. For initial experimentation, early implementations of this design will use the OID 1.3.6.1.4.1.44363.47.4 instead of id-pe-mtcCertificationAuthority-SHA256. Cloudflare has kindly donated the 1.3.6.1.4.1.44363.47 OID arc for use in this document. The fields of an MTCCertificationAuthority structure are defined as follows:

sigAlg is the CA cosigner's signature algorithm ().
minSerial and maxSerial describe the minimum and maximum allowed serial numbers from this CA, respectively. See for discussion on setting these values.

If this extension is present, the key described in subjectPublicKeyInfo is a CA cosigner key and subject to the usage restrictions described in . In particular, it MUST NOT be used to directly sign TBSCertificate structures. This extension indicates the subtree signature format defined in . If a later version of the protocol defines a new format, this SHOULD be represented in CA certificates with a new extension type. A CA certificate using this format SHOULD NOT be self-signed by the CA. Doing so would require writing the information in the issuance log. Instead, if used to represent a trust anchor, the certificate SHOULD be an unsigned certificate .

Certificates This section defines how to construct Merkle Tree Certificates, which are X.509 Certificates that assert the information in an issuance log entry.

Certificate Inputs A Merkle Tree Certificate is constructed from the following inputs:

A TBSCertificateLogEntry () contained in one of the CA's issuance logs ()
A subject public key whose hash matches the TBSCertificateLogEntry
The log_number and the zero-based entry index of that log entry within the issuance log, used to construct the certificate's serialNumber ().
An MTCProof () proving the entry's inclusion in a subtree, along with zero or more signatures () over that subtree, which together satisfy relying party requirements ()

By varying the choice of subtree and signatures, there can be multiple ways to prove the entry is in the log, and thus certified by the CA. defines how a certificate is constructed based on those choices. and define two profiles of Merkle Tree Certificates, standalone certificates and landmark-relative certificates, and how to select the subtree and signatures for them.

Certificate Format The information is encoded in an X.509 Certificate as follows: The TBSCertificate's version, issuer, validity, subject, issuerUniqueID, subjectUniqueID, and extensions MUST be equal to the corresponding fields of the TBSCertificateLogEntry. If any of issuerUniqueID, subjectUniqueID, or extensions is absent in the TBSCertificateLogEntry, the corresponding field MUST be absent in the TBSCertificate. Per , this means issuer MUST be the issuance log's CA ID as a PKIX distinguished name, as described in . The TBSCertificate's serialNumber is constructed from the zero-based index of the TBSCertificateLogEntry in the log and the log's number (). The serialNumber MUST be equal to (log_number << 48) | index. All serial numbers constructed in this way will be positive and at most 2⁶⁴-1. The TBSCertificate's subjectPublicKeyInfo contains the specified public key. Its algorithm field MUST match the TBSCertificateLogEntry's subjectPublicKeyAlgorithm. Its hash MUST match the TBSCertificateLogEntry's subjectPublicKeyInfoHash. The TBSCertificate's signature and the Certificate's signatureAlgorithm MUST contain an AlgorithmIdentifier whose algorithm is id-alg-mtcProof, defined below, and whose parameters is omitted. For initial experimentation, early implementations of this design will use the OID 1.3.6.1.4.1.44363.47.0 instead of id-alg-mtcProof. Cloudflare has kindly donated the 1.3.6.1.4.1.44363.47 OID arc for use in this document. The signatureValue contains an MTCProof structure, defined below using the TLS presentation language (): ; opaque HashValue[HASH_SIZE]; struct { TrustAnchorID cosigner_id; opaque signature<0..2^16-1>; } SubtreeSignature; struct { MTCLogEntryExtension extensions<0..2^16-1>; uint48 start; uint48 end; HashValue inclusion_proof<0..2^16-1>; SubtreeSignature signatures<0..2^24-1>; } MTCProof; ]]> extensions MUST contain the log entry's extensions value (). start and end MUST contain the corresponding parameters of the chosen subtree. inclusion_proof MUST contain a subtree inclusion proof () for the log entry and the subtree. signatures contains the chosen subtree signatures. In each signature, cosigner_id contains the cosigner ID () in its binary representation (), and signature contains the signature value as described in . The timestamp field used when computing the signature MUST be zero. Each element of the signatures field MUST have a unique cosigner_id. Elements MUST be ordered by cosigner_id (excluding length prefix) as follows:

Shorter byte strings are ordered before longer byte strings
Byte strings of the same length are ordered lexicographically

An MTCProof parser MUST reject the input if there are duplicate cosigner_id values, or if they are not ordered correctly. This can be done by checking each cosigner_id value comes strictly after the previous one in the above order. The MTCProof is encoded into the signatureValue with no additional ASN.1 wrapping. The most significant bit of the first octet of the signature value SHALL become the first bit of the bit string, and so on through the least significant bit of the last octet of the signature value, which SHALL become the last bit of the bit string.

Standalone Certificates A standalone certificate is a Merkle Tree certificate which contains sufficient signatures to allow a relying party to trust the choice of subtree, without any predistributed information beyond the cosigner(s) parameters. Standalone certificates can be issued without significant processing delay. When issuing a certificate, the CA first adds the TBSCertificateLogEntry to its issuance log. It then schedules a job to construct a checkpoint and collect cosignatures. The job proceeds as follows:

The CA signs the checkpoint with its key(s) ().
Using the procedure in , the CA determines the two subtrees that cover the entries added between this checkpoint and the most recent checkpoint.
The CA signs each subtree with its key(s) ().
The CA requests sufficient subtree cosignatures from external cosigners to meet relying party requirements (). Depending on the protocol for requesting subtree cosignatures (e.g. and ), this step may require first requesting a checkpoint cosignature () from each cosigner.
For each log entry in the interval, the CA constructs a certificate () from the inputs in , using the covering subtree and the subtree cosignatures collected in steps 3 and 4.

Step 4 is analogous to requesting SCTs from CT logs in Certificate Transparency, except that a single run of this job collects signatures for many certificates at once. The CA MAY request signatures from a redundant set of cosigners and select the ones that complete first. This document does not place any requirements on how frequently this job runs. More frequent runs result in lower issuance delay, but higher signing overhead. It is RECOMMENDED that CAs run at most one instance of this job at a time, starting the next instance after the previous one completes. A single run collects signatures for all entries since the most recent checkpoint, so there is little benefit to overlapping them. Less frequent runs may also aid relying parties that wish to directly audit signatures, as described in Section 5.2 of , though this document does not define such a system. This document does not prescribe the specific cosigner roles, or a particular protocol for requesting cosignatures. Protocols for cosigners can vary depending on the needs of that cosigner. Some example protocols are described in and . It is RECOMMENDED that the CA collect cosignatures for the authenticating party, but the authenticating party MAY collect additional cosignatures and add them to the certificate.

Landmark-Relative Certificates A landmark-relative certificate is a Merkle Tree certificate which contains no signatures and instead assumes the relying party had predistributed information about which subtrees were trusted. Landmark-relative certificates are an optional size optimization. They require a processing delay to construct, and only work in a sufficiently up-to-date relying party. Authenticating parties thus SHOULD deploy a corresponding standalone certificate alongside any landmark-relative certificate, and use some application-protocol-specific mechanism to select between the two. discusses such a mechanism for TLS .

Landmark Tree Sizes A CA that issues landmark-relative certificates MUST additionally maintain a landmark sequence. A landmark sequence is a sequence of landmarks, defined below. Landmarks are used as a common point of reference across the ecosystem for optimizing certificates. Each landmark consists of:

A landmark number, used as an identifier for the landmark
A tree size, which is the size of the tree at the time the landmark was allocated
An expiration time, represented as seconds since the Epoch (Section 4.19 of Volume 1 of )

The landmark sequence is append-only, with landmarks numbered consecutively from zero. Landmark zero MUST have a tree size of zero and an expiration of zero seconds since the Epoch. For each subsequent landmark, the tree size MUST be greater than that of the previous landmark, and the expiry MUST be greater or equal to that of the previous landmark. Each landmark has two landmark subtrees. The landmark subtrees for landmark number L as determined follows:

If L is zero, the landmark subtrees are [0, 0) and [0, 0).
Otherwise, let tree_size be landmark L's tree size and prev_tree_size be that of landmark L - 1.
The landmark subtrees are the two subtrees that cover [prev_tree_size, tree_size), as described in .

A landmark's expiration time MUST be greater or equal to the notAfter time of every TBSCertificateLogEntry whose index is less than the tree size. When allocating a landmark, CAs SHOULD set the expiration time to the current time plus the CA's maximum certificate lifetime. A landmark that is not yet expired is said to be active. Landmark zero is never active. At any time, a log's active landmark subtrees are the landmark subtrees of each currently active landmark. Active landmark subtrees are predistributed to the relying party as trusted subtrees, as described in . The above conditions imply that every unexpired entry in the log is either contained in some landmark subtree or was allocated sometime after the latest landmark. As the issuance log grows, CAs continuously allocate new landmarks. More frequent allocation reduces landmark-relative certificate delay, while less frequent allocation reduces the size of the relying party's predistributed state. As described in , relying parties maintain some upper bound on active landmarks per CA. CAs SHOULD allocate landmarks such that the number of active landmarks, across all their logs, is within the bound for supported relying parties. gives a RECOMMENDED procedure for allocating landmarks. Mistakes in landmark sequence allocation only impact availability, not security. That is, they will not cause the relying party to accept certificates for entries the CA did not certify. However, they might cause a relying party to reject some of the CA's otherwise valid landmark-relative certificates.

Allocating Landmarks It is RECOMMENDED that landmarks be allocated using the following procedure:

Let max_cert_lifetime by some upper bound on the CA's certificate lifetime.
Select some time_between_landmarks duration.
Define a series of consecutive, non-overlapping time intervals, each of duration time_between_landmarks.
At most once per time interval, run the following:
1. If the current log's tree size is equal to the its landmark's tree size, do nothing.
2. Otherwise, append a landmark to the current log whose tree size is the current tree size and whose expiry is the current time plus max_cert_lifetime.

This procedure ensures there are at most ceil(max_cert_lifetime / time_between_landmarks) + 1 active landmarks across all of the CA's logs. For example, if max_cert_lifetime is 7 days and time_between_landmarks is one hour, there will be at most 169 active landmarks, or 338 active landmark subtrees. The relying party state is then 10,816 bytes with SHA-256.

Publishing Landmarks The following format can be used to represent a CA's active landmarks. The format MUST contain the following sequence of lines. Each line MUST be terminated by a newline character (U+000A):

A header line consisting of a decimal integer, latest_landmark, with the landmark number of the CA's most recent landmark at the time of publishing. This value MUST be at most 2⁴⁸-1.
A sequence of num_active_landmarks + 1 lines, where num_active_landmarks is the number of active landmarks at the time of publishing. Decoders MUST reject documents where there are greater than latest_landmark such lines. Numbered consecutively from zero, line i in this sequence consists of:
- The tree size for landmark latest_landmark - i as a decimal integer. This value MUST be at most 2⁴⁸-1.
- A single space character (U+0020).
- The expiration time for landmark latest_landmark - i as a decimal integer containing seconds since the Epoch (Section 4.19 of Volume 1 of ).

Tree sizes MUST be strictly monotonically decreasing, and expiration times MUST be monotonically decreasing. There MUST be at least one expiration time before the current time. Decoders MUST reject documents that do not strictly conform to the above requirements, including extraneous whitespace and the lack of an expired landmark. A decoder MAY process only a prefix of this document, provided there is at least one expired landmark to denote the end of the active landmarks.

Constructing Landmark-Relative Certificates Given the inputs in and the corresponding log's landmark sequence, a landmark-relative certificate is constructed as follows:

Let idx be the entry index.
Let L be the lowest numbered landmark whose tree size is strictly greater than idx. If no such landmark has been allocated yet, wait for one to be allocated. If the entry has already expired and historical landmark information is unavoidable, abort the procedure.
Determine the L's subtrees () and select the unique one whose [start, end) interval contains idx.
Construct a certificate () using the selected subtree and no signatures.

The procedure above is not specific to the CA. Any party holding a standalone certificate () can construct the corresponding landmark-relative certificate by recovering the certificate inputs from it and obtaining the landmark sequence and inclusion proof hashes from the issuance log.

Size Estimates The inclusion proofs in standalone and landmark-relative certificates scale logarithmically with the size of the subtree. These sizes can be estimated with the CA's issuance rate. The byte counts below assume the issuance log's hash function is SHA-256. Some organizations have published statistics which can be used to estimate this rate for the Web PKI. As of September 18th, 2026:

reported around 682,000,000 active certificates for a single CA
reported around 2,900,000,000 unexpired certificates in CT logs, across all CAs
reported an issuance rate of around 591,000 certificates per hour, across all CAs

The current issuance rate across the Web PKI may not necessarily be representative of the Web PKI after a transition to short-lived certificates. Assuming a certificate lifetime of 7 days, and that subscribers will update their certificates 75% of the way through their lifetime (see ), every certificate will be reissued every 126 hours. This gives issuance rate estimates of around 5,400,000 certificates per hour and 23,000,000 certificates per hour, for the first two values above. Note the larger estimate is across all CAs, while subtrees would only span one CA. Using the per-CA short lifetime estimate, if the CA mints a checkpoint every 2 seconds, standalone certificate subtrees will span around 3,000 certificates, leading to 12 hashes in the inclusion proof, or 384 bytes. Standalone certificates additionally must carry a sufficient set of signatures to meet relying party requirements. If a new landmark is allocated every hour, landmark-relative certificate subtrees will span around 5,400,000 certificates, leading to 23 hashes in the inclusion proof, giving an inclusion proof size of 736 bytes, with no signatures. This is significantly smaller than a single ML-DSA-44 signature, 2,420 bytes, and almost ten times smaller than the three ML-DSA-44 signatures necessary to include post-quantum SCTs. Proof sizes grow logarithmically, so 32 hashes, or 1024 bytes, is sufficient for subtrees of up to 2³² (4,294,967,296) certificates.

Relying Parties This section discusses how relying parties verify Merkle Tree Certificates.

Relying Party Configuration In order to accept certificates from a Merkle Tree CA, a relying party MUST be configured with:

The CA's ID ()
The CA's log hash algorithm, e.g. SHA-256
The CA cosigner, and any other supported cosigners, as pairs of cosigner ID and public key
A policy on which combinations of cosigners to accept in a certificate ()
An optional list of trusted subtrees that are known to be consistent with the relying party's cosigner requirements ()
A list of revoked ranges of serial numbers ()

This information may be obtained from a CA certificate structure, defined in :

The CA ID is determined from the certificate's subject.
The log hash algorithm is determined from the type of the Merkle Tree CA extension.
The CA cosigner is determined from the certificate's subject public key and Merkle Tree CA extension. The CA's cosigner ID is the same as its CA ID. The relying party incorporates this cosigner into its cosigner policy based on the guidance in .
No trusted subtrees are directly represented by the CA certificate structure, but the relying party MAY incorporate trusted subtrees from out-of-band information.
The revoked serial number ranges include the half-open ranges [0, minSerial) and [maxSerial+1, 2^64), but the relying party MAY incorporate additional ranges from out-of-band information.

Verifying Certificate Signatures When verifying the signature of an X.509 certificate (Step (a)(1) of ) whose issuer is a Merkle Tree CA, the relying party performs the following procedure:

Check that the TBSCertificate's signature field is id-alg-mtcProof with omitted parameters. If this check fails, abort this process and fail verification.
Decode the signatureValue as an MTCProof, as described in . If decoding fails, including if signatureValue is not a multiple of 8 bits or has extra data after the MTCProof, abort this process and fail verification.
Let serial be the certificate's serial number. If serial is negative or greater than 2⁶⁴-1, abort this process and fail verification.
If serial is contained in one of the relying party's revoked ranges (), abort this process and fail verification.
Let index be the least significant 48 bits of serial and let log_number be serial >> 48. If log_number is zero, abort this process and fail verification.
Let log_id be the log ID constructed from the CA ID in issuer and the log_number ().
Construct a TBSCertificateLogEntry as follows:
1. Copy the version, issuer, validity, subject, issuerUniqueID, subjectUniqueID, and extensions fields from the TBSCertificate.
2. Set subjectPublicKeyAlgorithm to the algorithm field of the subjectPublicKeyInfo.
3. Set subjectPublicKeyInfoHash to the hash of the DER encoding of subjectPublicKeyInfo.
Construct an MTCLogEntry as follows:
1. Set type to tbs_cert_entry.
2. Set extensions to the MTCProof's extensions value.
3. Set tbs_cert_entry_data to the TBSCertificateLogEntry, encoded as described in .
Let entry_hash be the hash of the entry, MTH({entry}) = HASH(0x00 || entry), as defined in .
Let expected_subtree_hash be the result of evaluating the MTCProof's inclusion_proof for entry index, with hash entry_hash, of the subtree described by the MTCProof's start and end, following the procedure in . If evaluation fails, abort this process and fail verification.
If log_number, start, and end match a trusted subtree () for the CA, check that expected_subtree_hash is equal to the trusted subtree's hash. Return success if it matches and failure if it does not.
Otherwise, check that the MTCProof's signatures contain a sufficient set of valid signatures from cosigners to satisfy the relying party's cosigner requirements (). Unrecognized cosigners MUST be ignored. Signatures are verified as described in . For each signature verification, the CosignedMessage structure is constructed as follows:
1. Set the CosignedMessage's cosigner_name based on the cosigner ID as described in .
2. Set the CosignedMessage's timestamp to zero.
3. Set the CosignedMessage's log_origin based on log_id as described in .
4. Set the CosignedMessage's start and end to the MTCProof's start and end, respectively.
5. Set the CosignedMessage's subtree_hash to expected_subtree_hash.

This procedure only replaces the signature verification portion of X.509 path validation. The relying party MUST continue to perform other checks, such as checking expiry. In this procedure, entry_hash can equivalently be computed in a single pass from the DER-encoded TBSCertificate, without storing the full TBSCertificateLogEntry or MTCLogEntry in memory:

Initialize a hash instance.
Write the octet 0x00 to the hash. This is the domain separator for leaf nodes.
Write the extensions field from the MTCProof to the hash.
Write the big-endian, two-byte tbs_cert_entry value to the hash.
Write the TBSCertificate's version, issuer, validity, and subject fields to the hash.
Write the subjectPublicKeyInfo's algorithm field to the hash.
Write the octet 0x04 to the hash. This is an OCTET STRING identifier.
Write the octet L to the hash, where L is the hash length. (This assumes L is at most 127.)
Write H to the hash, where H is the hash of the entire subjectPublicKeyInfo field.
Write the remainder of the TBSCertificate contents octets to the hash, starting just after the subjectPublicKeyInfo field.
Finalize the hash and set entry_hash to the result.

This is possible because the structure in omits the TBSCertificateLogEntry's identifier and length octets.

Trusted Cosigners A relying party's cosigner policy determines the sets of cosigners that must sign a view of the issuance log before it is trusted. This document does not prescribe a particular policy, but gives general guidance. Relying parties MAY implement policies other than those described below, and MAY incorporate cosigners acting in roles not described in this document. In picking trusted cosigners, the relying party SHOULD ensure the following security properties:

Authenticity:: The relying party only accepts entries certified by the CA
Transparency:: The relying party only accepts entries that are publicly accessible, so that monitors, particularly the subject of the certificate, can notice any unauthorized certificates

Relying parties SHOULD ensure authenticity by requiring a signature from the CA cosigner key. This is analogous to the signature in a directly-signed X.509 certificate. If the relying party obtains CA information from a CA certificate, the CA cosigner key is determined as in . While a CA signature is sufficient to prove a subtree came from the CA, this is not enough to ensure the certificate is visible to monitors. A misbehaving CA might not operate the log correctly, either presenting inconsistent versions of the log to relying parties and monitors, or refusing to publish some entries. To mitigate this, relying parties SHOULD ensure transparency by requiring a quorum of signatures from additional cosigners. At minimum, these cosigners SHOULD enforce a consistent view of the log. For example, describes a lightweight "witness" cosigner role that checks this with consistency proofs. This is not sufficient to ensure durable logging. discusses mitigations for this. Alternatively, a relying party MAY require that cosigners serve a copy of the log, in addition to enforcing a consistent view. For example, describes a "mirror" cosigner role. Relying parties MAY accept the same set of additional cosigners across CAs. In applications that do not enforce transparency requirements, a relying party MAY implement a policy that only checks for a signature from the CA cosigner. This fits the pattern of many existing X.509 applications, where CA information is determined directly from a CA certificate, with no additional out-of-band information. Unrecognized cosignatures are ignored, so such applications can interoperate with certificates issued for transparency-enforcing applications that require additional cosigners. Cosigner roles are extensible without changes to certificate verification itself. Future specifications and individual deployments MAY define other cosigner roles to incorporate in relying party policies. discusses additional deployment considerations in cosigner selection.

Trusted Subtrees As an optional optimization, a relying party MAY incorporate a periodically updated, predistributed list of trusted subtrees from the CA. This allows the relying party to accept landmark-relative certificates () constructed against those subtrees. Each trusted subtree contains:

The log number of the containing log
The start and end values that define the subtree
The hash of the subtree

Trusted subtrees for a CA are determined by its active landmark subtrees, as described in . Before configuring the subtrees as trusted, the relying party MUST obtain assurance that each subtree is consistent with checkpoints observed by a sufficient set of cosigners (see ) to meet its cosigner requirements. It is not necessary that the cosigners have generated signatures over the specific subtrees, only that they are consistent. This criterion can be checked given:

Some reference checkpoint whose tree size is greater or equal to that of the latest landmark
For each cosigner, either:
- A cosignature on the reference checkpoint
- A cosigned checkpoint containing the referenced checkpoint and a valid Merkle consistency proof () between the two
For each subtree, a valid subtree consistency proof () between the subtree and the reference checkpoint

[[TODO: The subtree consistency proofs have many nodes in common. It is possible to define a single "bulk consistency proof" that verifies all the hashes at once, but it's a lot more complex.]] If a relying party is unable to validate some active landmark, it MAY discard that landmark, along with all landmarks in the log newer than it, while still using the older active landmarks that it was able to validate. For example, suppose the active landmarks have tree sizes 200, 300, 400, and 500, and the relying party was unable to validate any reference checkpoint of size 500 or higher. If the relying party is able to validate a reference checkpoint of size 350, it MAY incorporate subtrees from the first two landmarks. To bound local state, the relying party SHOULD define some upper bound on the number of active landmarks accepted per CA. If the CA exceeds this bound, the relying party SHOULD similarly discard the newest active landmarks to meet its limit. This document does not prescribe how relying parties obtain trusted subtrees. A relying party MAY, for example, use an application-specific update service, such as the services described in and . If the relying party considers the service sufficiently trusted (e.g. if the service provides the trust anchor list or certificate validation software), it MAY trust the update service to perform these checks. The relying party SHOULD incorporate its trusted subtree configuration in application-protocol-specific certificate selection mechanisms, to allow an authenticating party to select a landmark-relative certificate. The trust anchor IDs of the landmarks may be used as efficient identifiers in the application protocol. discusses how to do this in TLS .

Revoked Ranges For each supported Merkle Tree CA, the relying party maintains a list of revoked ranges of serial numbers. This can be used to revoke both ranges of entries in an issuance log and ranges of issuance logs, even if the contents are not known. When a relying party is first configured to trust an issuance log, it SHOULD be configured to revoke all serial numbers before the first available unexpired certificate at the time. This revocation SHOULD be periodically updated as entries expire. If using the format defined in , this can be configured with the minSerial value. This revocation allows the rest of a PKI to disregard old entries, even if they are not known to be expired. In particular:

A relying party could permit a CA to skip serving old entries (see ) if long-expired and revoked.
Newly-established monitors can skip processing long-expired and revoked entries.

A relying party with transparency requirements additionally SHOULD revoke all log numbers above some threshold to bound monitoring overhead. If using the format defined in , this can be configured with the maxSerial value. See . A misbehaving CA might correctly construct a globally consistent log, but refuse to make some entries or intermediate nodes available. Consistency proofs between checkpoints and subtrees would pass, but monitors cannot observe the entries themselves. Relying parties whose cosigner policies () do not require durable logging (e.g. via ) are particularly vulnerable to this. In this case, the indices of the missing entries will still be known, so relying parties can use this mechanism to revoke the unknown entries, possibly as an initial, targeted mitigation before complete CA removal. When a CA is found to be untrustworthy, relying parties SHOULD remove trust in that CA. To minimize the compatibility impact of this mitigation, index-based revocation can be used to only distrust entries after some index, while leaving existing entries accepted. This is analogous to the mechanism used in some PKIs. The revocation mechanism in this section is complementary to certificate-level revocation mechanisms. Because log entries are uniquely identified by their serial number and issuer, existing revocation mechanisms like CRLs and OCSP apply unchanged.

Use in TLS Most X.509 fields such as subjectPublicKeyInfo and X.509 extensions such as subjectAltName are unmodified in Merkle Tree certificates. They apply to TLS-based applications as in any X.509 certificate. The primary new considerations for use in TLS are:

Whether the authenticating party should send a certificate from one Merkle Tree CA, another Merkle Tree CA, or a directly-signing X.509 CA
Whether the authenticating party should send a standalone or landmark-relative certificate
What the relying party should communicate to the authenticating party to help it make this decision

Certificate selection in TLS, described in , incorporates both explicit relying-party-provided information in the ClientHello and CertificateRequest messages and implicit deployment-specific assumptions. This section describes a RECOMMENDED integration of Merkle Tree certificates into TLS trust anchor IDs (), but applications MAY use application-specific criteria in addition to, or instead of, this recommendation. Relying parties SHOULD NOT include Merkle Tree CAs in the certificate_authorities extension (). Doing so might inadvertently signal an unsupported landmark-relative certificate because they have the same issuer field as standalone certificates.

Standalone Certificates Authenticating and relying parties SHOULD use the trust_anchors extension to determine whether a standalone certificate would be acceptable. A standalone certificate has a trust anchor ID of the corresponding CA ID (). This trust anchor ID is additionally contained in the trust anchor groups defined in . CA IDs MAY be incorporated into other trust anchor groups, following the guidance in . [[TODO: Ideally we would negotiate cosigners. https://github.com/tlswg/tls-trust-anchor-ids/issues/54 has a sketch of how one might do this, though other designs are possible. Negotiating cosigners allows the ecosystem to manage cosigners efficiently, without needing to collect every possible cosignature and send them all at once. This is wasteful, particularly with post-quantum algorithms.]] A standalone certificate MAY also be sent without explicit relying party trust signals, however doing so means the authenticating party implicitly assumes the relying party trusts the issuing CA. This may be viable if, for example, the CA is relatively ubiquitous among supported relying parties.

Landmark-Relative Certificates An authenticating party SHOULD NOT send a landmark-relative certificate without a signal that the relying party trusts the corresponding landmark subtree. Even if the relying party is assumed to trust the issuing CA, the relying party may not have sufficiently up-to-date trusted subtrees. This can be represented with the trust_anchor_negotiation property in a CertificatePropertyList (see ), or other local configuration. TLS implementations SHOULD use the trust_anchors extension to determine this. A landmark-relative certificate issued by a CA with ID caID, log number N, and constructed from landmark L has a trust anchor ID of {caID landmarks(1) N L}. For example, the trust anchor ID for landmark 42 of CA 32473.100 and log number 8 is 32473.100.1.8.42. These trust anchor IDs are used when it is necessary to identify an individual landmark, e.g. as in the recovery mechanism described in . To more efficiently express a relying party's complete landmark state, these IDs are contained in trust anchor groups defined in , which allow relying parties to express their landmark state with a single ID. If both a landmark-relative and a standalone certificate are usable, an authenticating party SHOULD preferentially use the landmark-relative certificate. A landmark-relative certificate asserts the same information as its standalone counterpart, but is expected to be smaller.

Single-Log Landmark Groups Relying parties support many landmarks per log at a time. To compactly represent this, each log ID implicitly defines a series of trust anchor groups () called landmark groups. For each Merkle Tree Certificates CA with ID caID, each log number N, and each landmark number L, the ID {caID landmarkGroups(2) N L} defines a landmark group. It contains the following trust anchor IDs:

caID itself (see ). This selects all standalone certificates.
{caID landmarks(1) N M} for all M from 0 to L, inclusive. This selects landmark-relative certificates from active landmarks up to L.

To support these groups in the authenticating party, CAs SHOULD configure certificates to match the following trust anchor groups (Sections and of ):

A standalone certificate SHOULD include a trust anchor ID pattern of caID.2.{0-}.{0-}.
A landmark-relative log number N and landmark L SHOULD include a trust anchor ID pattern of caID.2.N.{L-}.

For example, suppose a CA with ID 32473.100 issues a certificate in landmark 42 of log 8:

The standalone certificate has a trust anchor ID of 32473.100 and is contained in groups 32473.100.2.{0-}.{0-}.
The landmark-relative certificate has a trust anchor ID of 32473.100.1.8.42 and is contained in groups 32473.100.2.8.{42-}.

A relying party whose latest trusted subtree () in log N is landmark L SHOULD configure the trust_anchors extension to advertise the above landmark group. This signals support for both standalone certificates and supported landmarks. For example, a relying party which is up-to-date as of landmark 42 of log 8 of CA 32473.100 would send an ID of 32473.100.2.8.42. This would signal the following certificates:

Any standalone certificate from 32473.100, no matter the log or landmark number.
Any landmark-relative certificate from 32473.100 from landmarks 23 through 42, inclusive, of log 8.

If this landmark information becomes too stale, such a relying party SHOULD switch to advertising just the CA ID. In the above example, this would be 32473.100.

Timestamped Landmark Groups Landmark groups for a single CA, described above, allow relying parties to advertise one ID per supported CA. Depending on the number of trust anchors, this can be sufficient to efficiently represent relying party state. When needed, describes how PKIs can use trust anchor groups that span multiple CAs. This section defines a variation of the versioning construction described in , as applied to landmarks. Trust anchor groups containing Merkle Tree CAs can represent landmarks with an OID component based on a predictable clock. Concretely, the family of groups is parameterized by:

A base OID arc base
A timestamp start_time
A time duration tick_duration

Given non-negative integers V and T, the group base.V contains standalone certificates issued by some CA in version V of the group. The group base.V.T contains:

Standalone certificates issued by some CA in version V of the group.
Landmark-relative certificates issued one of the above CAs, provided the landmark was active at time start_time + T * tick_duration.

start_time SHOULD be set to sometime before the group is in use. tick_duration SHOULD be set near the expected time between landmarks in the group, e.g. one hour. This predictable cadence allows the CA to describe the trust anchor groups () for issued certificates without additional coordination. Concretely, if a CA was added in V_min, was removed in V_max + 1, and issues a certificate whose landmark was first active at time T_min and last active at time T_max:

The standalone certificate is contained in groups base.{V_min-V_max} and base.{V_min-V_max}.{0-}.
The landmark-relative certificate is contained in groups base.{V_min-V_max}.{T_min-T_max}

If the CA has not been removed in the latest version, V_max is infinity, similar to the construction described in . T_min and T_max are measured based on start_time and tick_duration as described above. A relying party sets V based on its current trust anchors and T based on the age of its landmark information. If its landmarks are too stale, it sends base.V without any landmark timestamp. In some cases, the relying party's landmark information may only be partially up-to-date. The relying party, or its update service, may be unable to reach one CA in the group, e.g. due to a transient outage. This complicates timestamp-based strategies:

If the relying party uses an older timestamp, it will not signal its up-to-date state for the reachable CAs. This means a single unreachable CA can disrupt service for certificates issued by unrelated CAs.
If the relying party uses a newer timestamp, the relying party may signal support for landmarks it does not have. This risks connection failures. If the unreachable CA issued recent landmark-relative certificates, those certificates will fail validation.

The relying party can mitigate this in a number of ways:

If the trust anchor group consists of CAs from the same operator, waiting until all CAs are reachable will be minimally disruptive.
The relying party can opt to send the group with an older timestamp, combined with other, smaller groups at newer timestamps to better describe its state.
A client relying party can send the newer timestamp and, in the event the unreachable CA did issue recent landmark-relative certificates, rely on the recovery mechanism described in to recover from any signaling failures.

ACME Extensions This section describes how to issue Merkle Tree certificates using ACME .

Optional Certificates describes how an ACME server uses the "alternate" link relation to serve multiple certificate chains for an ACME order. An ACME client might fetch all of them and deploy them in the authenticating party. Different relying parties need different chains, so the ACME client might reasonably treat any unavailable alternate as an error. This behavior is not ideal for a landmark-relative certificate, which is available asynchronously and is not intended to delay the corresponding standalone certificate. This section defines the "acme-optional-alternate" link relation. When serving a certificate, an ACME server MAY provide one or more link relation header fields of type "acme-optional-alternate". "acme-optional-alternate" identifies an alternate certificate chain, but one that is optional. Relying parties that accept the optional alternate are expected to also accept either the original certificate chain or chains served under the "alternate" link relation. If the certificate chain is not yet available, the "acme-optional-alternate" URL SHOULD serve an HTTP 202 (Accepted) response, with a Retry-After header () estimating when it will become available. An ACME client MAY fetch these URLs to collect additional alternate certificate chains. If the resource is unavailable, the ACME client SHOULD NOT fail the overall transaction. If the resource returns an HTTP 202 (Accepted) response, the ACME client SHOULD retry the request later, according to the Retry-After header, but this process SHOULD be independent of deploying other chains in the ACME order. In particular, if deploying a new service, the ACME client SHOULD NOT block deployment on optional alternates. If renewing certificates, the ACME client MAY opt to wait for optional alternates to simplify certificate replacement, but only while the previous certificates remain valid.

Using ACME with Merkle Tree Certificates Standalone and landmark-relative certificates represent a single issuance event, so they are returned from the same order. When processing an order for a Merkle Tree certificate, the ACME server moves the order to the "valid" state after the standalone certificate is available. The order's certificate URL then serves the standalone certificate, constructed as described in . The standalone certificate response SHOULD additionally carry an "acme-optional-alternate" URL for the landmark-relative certificate. The landmark-relative certificate will typically not yet be available, so it initially serves an HTTP 202 response, as described in . Once the next landmark is allocated, the ACME server constructs a landmark-relative certificate, as described in , and serves it from the "acme-optional-alternate" URL. When downloading either certificate (), ACME clients supporting Merkle Tree certificates SHOULD send "application/pem-certificate-chain-with-properties" in their Accept header (). ACME servers issuing Merkle Tree certificates SHOULD then respond with that content type to include a CertificatePropertyList. The CertificatePropertyList SHOULD include trust anchor ID information as described in . describes the trust anchor ID assignments for standalone and landmark-relative certificates. At minimum, the ACME server SHOULD include:

A trust_anchor_id property with the trust anchor IDs described in and
A trust_anchor_groups property with the information described in

If the CA participates in other landmark groups, e.g. , the ACME server SHOULD include the corresponding information in trust_anchor_groups. The ACME server SHOULD include a trust_anchor_negotiation property with the landmark-relative certificate. This indicates the landmark-relative certificate requires a trust anchor ID match to indicate that the relying party recognizes the landmark. The ACME server MAY include or omit trust_anchor_negotiation with the standalone certificate, based on the criteria described in Sections and of .

Deployment Considerations

Operational Costs

Certification Authority Costs While Merkle Tree certificates expect CAs to operate logs, the costs of these logs are expected to be much lower than a CT log from or : does not constrain the API to the one defined in or . If the PKI uses a tile-based protocol, such as (profiled for Merkle Tree Certificates in ), the issuance log benefits from the improved caching properties of such designs. Unlike a CT log, an issuance log does not have public submission APIs. Log entries are only added by the CA directly. Costs are thus expected to scale with the CA's own issuance. A CA only needs to produce a digital signature for every checkpoint, rather than for every certificate. The lower signature rate requirements could allow more secure and/or economical key storage choices. Individual entries are kept small and do not scale with public key or signature sizes. This mitigates growth from post-quantum algorithms. Public keys in entries are replaced with fixed-sized hashes. There are no signatures in entries themselves, and only signatures on the very latest checkpoint are retained. Every new checkpoint completely subsumes the old checkpoint, so there is no need to retain older signatures. Likewise, a subtree is only signed if contained in another signed checkpoint. Explicit revocation of old entries () allows a long-lived log to serve only the more recent entries, scaling with the size of the retention window, rather than the log's total lifetime. Mirrors of the log can also reduce CA bandwidth costs, because monitors can fetch data from mirrors instead of CAs directly. In PKIs that deploy mirrors as part of cosigner policies, relying parties could set few availability requirements on CAs, as described in .

Cosigner Costs The costs of cosigners vary by cosigner role. A consistency-checking cosigner, such as , requires very little state and can be run with low cost. A mirroring cosigner, such as , performs a role comparable to CT logs, but several of the cost-saving properties in also apply: improved protocols, smaller entries, less frequent signatures, and partial log serving. While a mirror does need to accommodate another party's (the CA's) growth rate, it grows only from new issuances from that one CA. If one CA's issuance rate exceeds the mirror's capacity, that does not impact the mirror's copies of other CAs. Mirrors also do not need to defend against a client uploading a large number of existing certificates all at once. Submissions are naturally batched and serialized.

Monitor Costs In a CT-based PKI, every log carries a potentially distinct subset of active certificates. Monitors must check the contents of every CT log. At the same time, certificates are commonly synchronized between CT logs. As a result, a monitor will typically download each certificate multiple times, once for every log. In Merkle Tree Certificates, each entry appears in exactly one log. A relying party might require a log to be covered by a quorum of mirrors, but each mirror is cryptographically verified to serve the same contents. Once a monitor has obtained some entry from one mirror, it does not need to download it from the others. In addition to downloading each entry only once, the entries themselves are smaller, as discussed in .

Choosing Cosigners In selecting trusted cosigners and cosigner requirements (), relying parties navigate a number of trade-offs: A consistency-checking cosigner, such as , is inexpensive to run, but does not guarantee durable logging. A mirroring cosigner is more expensive and may take longer to cosign structures. Requiring a mirror signature provides stronger guarantees to the relying party, which in turn can reduce the requirements on CAs (see ), however it may cause certificate issuance to take longer. That said, mirrors are comparable to CT logs, if not cheaper (see ), so they may be appropriate in PKIs where running CT logs is already viable. Relying parties that require larger quorums of trusted cosigners can reduce the trust placed in any individual cosigner. However, larger quorums result in larger, more expensive standalone certificates. The cost of standalone certificates will depend on how frequently the landmark optimization occurs in a given PKI. Conversely, relying parties that require smaller quorums have smaller standalone certificates, but place more trust in their cosigners. Relying party policies also impact monitor operation. If a relying party accepts any one of three cosigners, monitors SHOULD check the checkpoints of all three. Otherwise, a malicious CA may send different split views to different cosigners. More generally, monitors SHOULD check the checkpoints in the union of all cosigners trusted by all supported relying parties. This is an efficient check because, if the CA is operating correctly, all cosigners will observe the same tree. Thus the monitor only needs to check consistency proofs between the checkpoints, and check the log contents themselves once. Monitors MAY also rely on other parties in the transparency ecosystem to perform this check.

Log Availability CAs and mirrors are expected to serve their log contents over HTTP. It is possible for the contents to be unavailable, either due to temporary service outage or because the log does not serve long-expired entries. If some resources are unavailable, they may not be visible to monitors. As in CT, PKIs that deploy Merkle Tree certificates SHOULD establish availability policies. These policies SHOULD be adhered to by trusted CAs and mirrors, and enforced by relying party vendors as a condition of trust. Exact availability policies for these services are out of scope for this document, but this section provides some general guidance. Availability policies MAY permit CAs and mirrors to stop serving old, long-expired entries. If so, such policies SHOULD, at minimum, require CAs and mirrors to retain entries until they have been revoked in up-to-date relying parties. See for details. This is analogous to the CT practice of temporal sharding , except the issuance log remains compatible with older, unupdated relying parties. PKIs that require mirror cosignatures () can impose minimal to no availability requirements on CAs without compromising transparency goals. If a CA never makes an entry available, mirrors will be unable to update. This will prevent relying parties from accepting the undisclosed entries. However, a CA that is persistently unavailable may not offer sufficient benefit to be used by authenticating parties or trusted by relying parties. However, if a mirror's interface becomes unavailable, monitors may be unable to check for unauthorized issuance, if the entries are not available in another mirror. This does compromise transparency goals. As such, availability policies SHOULD set availability expectations on mirrors. This can also be mitigated by using multiple mirrors, either directly enforced in cosigner requirements, or by keeping mirrors up-to-date with each other. In PKIs that do not require mirroring cosigners, the CA's serving endpoint is more crucial for monitors. Such PKIs SHOULD set availability requirements on CAs. In each of these cases, the serial numbers of unavailable entries are known. Availability failures can thus be mitigated by revocation, as described in , likely as a first step in a broader distrust.

Certificate Renewal When an authenticating party requests a certificate, the landmark-relative certificate will not be available until the next landmark is ready. From there, the landmark-relative certificate will not be available until relying parties receive new trusted subtrees. To maximize coverage of landmark-relative certificates, authenticating parties performing routine renewal SHOULD request a new Merkle Tree certificate before the previous Merkle Tree certificate expires. Renewing around 75% of the way through the previous certificate's lifetime is RECOMMENDED. Authenticating parties additionally SHOULD retain both the new and old certificates in the certificate set until the old certificate expires. As the new subtrees are delivered to relying parties, certificate negotiation will transition relying parties to the new certificate, while retaining the old certificate for relying parties that are not yet updated. The above also applies if the authenticating party is performing a routine key rotation alongside the routine renewal. In this case, certificate negotiation would pick the key as part of the certificate selection. This slightly increases the lifetime of the old key but maintains the size optimization continuously. If the service is rotating keys in response to a key compromise, this option is not appropriate. Instead, the service SHOULD immediately discard the old key and request a standalone certificate and the revocation of the previous certificate. This will interrupt the size optimization until the new landmark-relative certificate is available and relying parties are updated.

Privacy Considerations The Privacy Considerations described in apply to their use with Merkle Tree Certificates. In particular, relying parties that share an update process for trusted subtrees () will fetch the same stream of updates. However, updates may reach different users at different times, resulting in some variation across users. This variation may contribute to a fingerprinting attack . If the Merkle Tree CA trust anchors are sent unconditionally in trust_anchors, this variation will be passively observable. If they are sent conditionally, e.g. gated on the recovery flow, the trust anchor list will require active probing.

Security Considerations

Authenticity A key security requirement of any PKI scheme is that relying parties only accept assertions that were certified by a trusted certification authority. Merkle Tree certificates achieve this by ensuring the relying party only accepts authentic subtree hashes:

In standalone certificates, the relying party's cosigner requirements () are expected to include some signature by the CA's cosigner. The CA's cosigner () is defined to certify the contents of every checkpoint and subtree that it signs.
In landmark-relative certificates, the cosigner requirements are checked ahead of time, when the trusted subtrees are predistributed ().

Given a subtree hash computed over entries that the CA certified, it must be computationally infeasible to construct an entry not on this list, and an inclusion proof, such that inclusion proof verification succeeds. This requires using a collision-resistant hash in the Merkle Tree construction. The subject public key is itself also hashed before incorporating into the log. This hash depends on second-preimage resistance. To authorize the wrong public key for some existing log entry (e.g. one that describes a target's identity), an attacker must find some other public key with the same hash. While an attacker able to compute collisions might find two public keys with the same hash, that hash will not be in any existing log entry. The attacker would need to be authorized to request certification of a new entry with this hash. Such an attacker could only certify colliding pairs of public keys for its own identities.

Transparency The transparency mechanisms in this document do not prevent a CA from issuing an unauthorized certificate. Rather, they provide comparable security properties as Certificate Transparency in ensuring that all certificates are either rejected by relying parties, or visible to monitors and, in particular, the subject of the certificate. Compared to Certificate Transparency, some of the responsibilities of a log have moved to the CA. All signatures generated by the CA in this system are assertions about some view of the CA's issuance log. However, a CA does not need to function correctly to ensure transparency properties. Relying parties are expected to require a quorum of additional cosigners, which together enforce properties of the log () and prevent or detect CA misbehavior: A CA might violate the append-only property of its log and present different views to different parties. However, each individual cosigner will only follow a single append-only view of the log history. Provided the cosigners are correctly operated, relying parties and monitors will observe consistent views. Views that were not cosigned at all may not be detected, but they also will not be accepted by relying parties. If the CA sends one view to some cosigners and another view to other cosigners, it is possible that multiple views will be accepted by relying parties. However, in that case monitors will observe that cosigners do not match each other. Relying parties can then react by revoking the range of inconsistent serials (), and likely removing the CA. If the cosigners are mirrors, the underlying entries in both views will also be visible. A CA might correctly construct its log, but refuse to serve some unauthorized entry. The impact depends on the relying party's cosigner policy:

If the relying party requires cosignatures from trusted mirrors, the entry will either be visible to monitors in the mirrors, or have never reached a mirror. In the latter case, the entry will not have been cosigned, so the relying party would not accept it.
If the relying party accepts log views without a trusted mirror, the unauthorized entry may not be available. However, the existence of some entry at that index will be visible, so monitors will know the CA is failing to present an entry. This is sufficient to determine the serial number, so relying parties can then react by revoking the undisclosed entries (), and likely removing the CA.

Log Failures Merkle Tree Certificates introduce additional state to PKI deployments and thus new kinds of operational failures. CAs are required to only sign subtree hashes that are consistent with a single append-only view of each issuance log. A CA might violate this as a result of operational failures. For example:

A CA loses some state and signs subtree hashes from two inconsistent copies of the log
A CA miscalculates some hash and signs a subtree hash that cannot be computed from some underlying sequence of entries

As described in , PKIs can use additional cosigners to provide transparency guarantees even in the face of such CA violations. In doing so, individual cosigners may be locked to only one of two views of the log or unable to sign further checkpoints because some hash's preimage is unknown. It may then no longer be possible to add entries to the log that are trusted by existing relying parties. Whether by accident or compromise, these violations are ultimately CA failures. However, it is useful for the CA instance to remain functional during and after incident management:

While the incident is diagnosed, authenticating parties may still need new certificates.
If relying parties consider the CA operator and the CA instance still trustworthy, repairing the incident without changing the CA requires less overhead.
If relying parties consider either the CA operator or the CA instance no longer trustworthy and in need of replacement, the CA may still be needed to serve older, unupdated relying parties.

This is mitigated by a CA instance consisting of a series of issuance logs (). After a log failure, the CA SHOULD increment its current issuance log to restore availability. Both the underlying log failure and the use of a new issuance log will be visible to monitors and SHOULD be treated as a PKI incident. Such PKI incidents can be handled by some combination of:

Revoking the diverging log indices ()
Reevaluating trusted CAs and, if necessary, removing the old CA instance and switching to a new CA instance

In the latter case, the CA operator MAY continue to operate the removed CA instance if, for example, there remain unupdated relying parties that require it.

Limiting Issuance Logs While multiple issuance logs help mitigate log failures, as described in , they introduce transparency risks. If a CA violates the requirement to only use one issuance log at a time, it might add an entry in some far future log number. To be accepted in transparency-enforcing relying parties, the log state must still be cosigned. However, monitors may not know which log numbers to monitor. PKIs with transparency requirements SHOULD mitigate this by only accepting a limited range of log numbers in relying parties, transparency cosigners, or both. This limit MAY be set to a fixed value or a rolling value that is updated whenever the CA switches its current log. Fixed values require committing to a limit of recoverable log failures over the lifetime of a CA. Log number limits in relying parties can be implemented by revoking all serial numbers above some threshold. (See .) If using the format described in , this can be implemented with the maxSerial field.

Public Key Hashes Unlike Certificate Transparency, the mechanisms in this document do not provide the subject public keys, only the hashed values. This is intended to reduce log serving costs, particularly with large post-quantum keys. As a result, monitors look for unrecognized hashes instead of unrecognized keys. Any unrecognized hash, even if the preimage is unknown, indicates an unauthorized certificate. This optimization complicates studies of weak public keys, e.g. . Such studies will have to retrieve the public keys separately, such as by connecting to the TLS servers, or fetching from the CA if it retains the unhashed key. This document does not define a mechanism for doing this, or require that CAs or mirrors retain unhashed keys. The transparency mechanisms in this protocol are primarily intended to allow monitors to observe certificate issuance.

Non-Repudiation When a monitor finds an unauthorized certificate issuance in a log or mirror, it must be possible to prove the CA indeed certified the information in the entry. However, only the latest signed checkpoint may be retained by the transparency ecosystem, so it may not be possible to reconstruct the exact certificate seen by relying parties. However, per , any subtree signature is a binding assertion by the CA that it has certified every entry in the subtree. Thus, given any signed checkpoint that contains the unauthorized entry, a Merkle inclusion proof () is sufficient to prove the CA issued the entry. This is analogous to how, in , CAs are held accountable for signed CT precertificates. The transparency ecosystem does not retain unhashed public keys, so it also may not be possible to construct a complete certificate from the signed checkpoint and inclusion proof. However, if the log entry's subjectPublicKeyInfoHash does not correspond to an authorized key for the subject of the certificate, the entry is still unauthorized. A Merkle Tree CA is held responsible for all log entries it certifies, whether or not the preimage of the hash is known.

Extensibility MTCLogEntry () contains several extension points:

New X.509 extensions can be added to TBSCertificateLogEntry.
New MTCLogEntryType values define new formats for the entry contents.
New MTCLogEntryExtensionType values define new entry extension fields.

X.509 extensions apply to Merkle Tree Certificates without any modifications. The two entry-level extension points are new to this protocol. Older CAs, cosigners, relying parties, and monitors may encounter unrecognized entries: Different cosigner roles interact with extensions differently. Some roles, e.g. and , do not interpret entry contents. Unrecognized extensions do not impact these roles. Other roles, such as CA cosigners, have semantics that depend on the entry contents. If a cosigner role interprets log entry contents, it MUST define how it interacts with unrecognized types and extensions. forbids a CA from logging or signing entries that it does not recognize. A CA cannot faithfully claim to certify information if it does not understand it. This is analogous to how a correctly-operated X.509 CA can never sign an unrecognized X.509 extension. Unrecognized entry types do not impact older relying parties. In , the relying party constructs the MTCLogEntry that it expects. The unrecognized entry will have a different type value, so the proof will never succeed, assuming the underlying hash function remains collision-resistant. However, unrecognized entry extensions will be ignored by relying parties, analogously to a non-critical X.509 extension. Entry extensions thus SHOULD be defined so that this is safe. If a monitor observes an entry with unknown type or entry extension, it may not be able to determine if it is of interest. For example, it may be unable to tell whether it covers some relevant DNS name. Until the monitor is updated to reflect the current state of the PKI, the monitor may be unable to detect all misissued certificates. This situation is analogous to the addition of a new X.509 extension. When relying parties add support for log entry types or new X.509 extensions, they SHOULD coordinate with monitors to ensure the transparency ecosystem is able to monitor the new formats.

Certificate Malleability An ASN.1 structure like X.509's Certificate is an abstract data type that is independent of its serialization. There are multiple encoding rules for ASN.1. Commonly, protocols use DER , such as . This aligns with , which says X.509 signatures are computed over the DER-encoded TBSCertificate. After signature verification, applications can assume the DER-encoded TBSCertificate is not malleable. When the signature verification process in first transforms the TBSCertificate into a TBSCertificateLogEntry, it preserves this non-malleability. There is a unique valid DER encoding for every abstract TBSCertificate structure, so malleability of the DER-encoded TBSCertificate reduces to malleability of the TBSCertificate value:

The version, issuer, validity, subject, issuerUniqueID, subjectUniqueID, and extensions fields are copied from the TBSCertificate to the TBSCertificateLogEntry unmodified, so they are directly authenticated by the inclusion proof.
serialNumber is omitted from TBSCertificateLogEntry, but its value determines the inclusion proof index, which authenticates it.
The redundant signature field in TBSCertificate is omitted from TBSCertificateLogEntry, but checks for an exact value, so no other values are possible.
subjectPublicKeyInfo is hashed as subjectPublicKeyInfoHash in TBSCertificateLogEntry. Provided the underlying hash function is collision-resistant, no other values are possible for a given log entry.

X.509 implementations often implement by equivalently retaining the original received DER encoding, rather than recomputing the canonical DER encoding TBSCertificate. This optimization is compatible with the assumptions above. Some non-conforming X.509 implementations use a BER parser instead of DER, and then apply this optimization to the received BER encoding. BER encoding is not unique, so this does not produce the same result. In such implementations, the BER-encoded TBSCertificate becomes also non-malleable, and applications may rely on this. To preserve this property in Merkle Tree Certificates, such non-conforming implementations MUST do the following when implementing :

Reparse the initial identifier (the SEQUENCE tag) and length octets of the TBSCertificate structure with a conforming DER parser and fail verification if invalid.
When copying the version, issuer, validity, subject, issuerUniqueID, subjectUniqueID, and extensions fields, either copy over the observed BER encodings, or reparse each field with a conforming DER parser and fail verification if invalid.
Reparse the serialNumber field with a conforming DER parser and fail verification if invalid.
Reparse the signature field with a conforming DER parser and fail verification if invalid. Equivalently, check for an exact equality with the expected, DER-encoded value.
When hashing subjectPublicKeyInfo, either hash the observed BER encoding, or reparse the structure with a conforming DER parser and fail verification if invalid.

These additional checks are redundant in X.509 implementations that use a conforming DER parser. requires that the TBSCertificateLogEntry in an MTCLogEntry be DER-encoded, so applying a stricter parser will be compatible with conforming CAs. While these existing non-conforming implementations may be unable to switch to a DER parser due to compatibility concerns, Merkle Tree Certificates are new, so there is no existing deployment of malformed BER-encoded TBSCertificateLogEntry structures. The above only ensures the TBSCertificate portion is non-malleable. In Merkle Tree Certificates, similar to an ECDSA X.509 signature, the signature value is malleable. Multiple MTCProof structures may prove a single TBSCertificate structure. Additionally, in all X.509-based protocols, a BER-based parser for the outer, unsigned Certificate structure will admit malleability in those portions of the encoding. Applications that derive a unique identifier from the Certificate MUST instead use the TBSCertificate, or some portion of it, for Merkle Tree Certificates.

Revocation This document does not define a new certificate-level revocation mechanism. Existing mechanisms like CRLs and OCSP apply unchanged to Merkle Tree certificates. The sequential serial numbers assigned by issuance logs may enable future improvements to revocation, but such work is out of scope for this document.

Signature Domain Separation The signature format defined in includes a fixed label prefix to ensure domain separation. Provided other uses of the same key use a non-overlapping prefix, signatures in one context cannot be substituted for those in another. permits a CA cosigner key to be used to sign CRLs and OCSP responses. These signatures do not include a domain separation prefix. Instead, X.509 relies on an undocumented assumption that the TBSCertificate, TBSCertList, and OCSP ResponseData structures do not overlap at the level of individual ASN.1 fields. These ASN.1 structures all begin with a SEQUENCE tag, which is encoded in DER as 0x30 or the ASCII digit "0". The domain separation label used in , subtree/v1\n\0, does not begin with "0", so their inputs do not overlap. More generally, this label is not a prefix of any DER or BER encoding. Domain separation analysis based on the structures themselves is fragile, particularly when individual ASN.1 fields must be analyzed. This document depends on a structure-level analysis for CRLs and OCSP responses due to how these legacy protocols were defined. Future uses of the key SHOULD use a more robust mechanism, namely a fixed label prefix or a context string parameter if the signature scheme supports it.

Subordinate Certification Authorities Merkle Tree Certificates' transparency properties only apply to certificates directly issued by the CA, not certification paths. The CA might issue a certificate that describes an unconstrained, subordinate, non-MTC CA. Certificates issued by the subordinate CA would not be visible in the MTC CA's issuance log and thus may not be visible to monitors. However, the subordinate CA certificate that enables this bypass will still be visible in the issuance logs. Although the scope is larger, this scenario is similar to an unauthorized end-entity certificate and can be handled analogously: Relying parties with transparency requirements SHOULD define policy requirements on trusted CAs that prevent these bypasses, with any violation treated as an unauthorized certificate. For example, a relying party might require that all subordinate CAs have name constraints () or forbid subordinate CAs entirely. In addition to holding CAs responsible for meeting these policies, relying parties SHOULD programmatically enforce these policies as part of certification path validation. Monitors SHOULD monitor for adherence to applicable policies as part of monitoring for unauthorized certificates. For example, a monitor that looks for entries covering example.com SHOULD look for either a subject alternative name () of example.com or a basic constraints () extension with the cA boolean set to true. It is not sufficient to constrain the MTC CA with a path length constraint () of zero. Self-issued certificates do not contribute to path length constraints, so such an MTC CA might still issue CA certificates with the same name as itself.

IANA Considerations

Module Identifier IANA is requested to add the following entry in the "SMI Security for PKIX Module Identifier" registry :

Decimal	Description	References
TBD	id-mod-mtc-2025	[this-RFC]

Algorithm IANA is requested to add the following entry to the "SMI Security for PKIX Algorithms" registry :

Decimal	Description	References
TBD	id-alg-mtcProof	[this-RFC]

Certificate Extension IANA is requested to add the following entry to the "SMI Security for PKIX Certificate Extension" registry :

Decimal	Description	References
TBD	id-pe-mtcCertificationAuthority-SHA256	[this-RFC]

Relative Distinguished Name Attribute IANA is requested to add the following entry to the "SMI Security for PKIX Relative Distinguished Name Attribute" registry :

Decimal	Description	References
TBD	id-rdna-trustAnchorID	[this-RFC]

Link Relation Type IANA is requested to add the following entry to the "Link Relation Types" registry :

Relation Name:: acme-optional-alternate
Description:: Refers to an optional alternate certificate chain, which may not be available immediately. Relying parties that accept the alternate are expected to also accept the original certificate, so it is not an error if the alternate is unavailable.
Reference:: [this-RFC],

References Normative References Information technology - ASN.1 encoding rules: Specification of Basic Encoding Rules (BER), Canonical Encoding Rules (CER) and Distinguished Encoding Rules (DER) ITU-T New ASN.1 Modules for the Public Key Infrastructure Using X.509 (PKIX) The Public Key Infrastructure using X.509 (PKIX) certificate format, and many associated formats, are expressed using ASN.1. The current ASN.1 modules conform to the 1988 version of ASN.1. This document updates those ASN.1 modules to conform to the 2002 version of ASN.1. There are no bits-on-the-wire changes to any of the formats; this is simply a change to the syntax. This document is not an Internet Standards Track specification; it is published for informational purposes. TLS Trust Anchor Identifiers OpenSSL Google LLC Meta This document defines the TLS Trust Anchors extension, a mechanism for a TLS client or server to select a certificate to present based on the peer's trusted certification authorities. It describes certification authorities more succinctly than the TLS Certificate Authorities extension. Key words for use in RFCs to Indicate Requirement Levels In many standards track documents several words are used to signify the requirements in the specification. These words are often capitalized. This document defines these words as they should be interpreted in IETF documents. This document specifies an Internet Best Current Practices for the Internet Community, and requests discussion and suggestions for improvements. Ambiguity of Uppercase vs Lowercase in RFC 2119 Key Words RFC 2119 specifies common key words that may be used in protocol specifications. This document aims to reduce the ambiguity by clarifying that only UPPERCASE usage of the key words have the defined special meanings. The Transport Layer Security (TLS) Protocol Version 1.3 This document specifies version 1.3 of the Transport Layer Security (TLS) protocol. TLS allows client/server applications to communicate over the Internet in a way that is designed to prevent eavesdropping, tampering, and message forgery. This document obsoletes RFC 8446, which specified TLS 1.3. This document obsoletes RFC 5246 (specifying TLS 1.2) and RFCs 5077, 6961, 7627, and 8422, all of which pertain to TLS 1.2 or earlier, and updates RFCs 5705 and 6066. This document also specifies new requirements for TLS 1.2 implementations. Certificate Transparency Version 2.0 This document describes version 2.0 of the Certificate Transparency (CT) protocol for publicly logging the existence of Transport Layer Security (TLS) server certificates as they are issued or observed, in a manner that allows anyone to audit certification authority (CA) activity and notice the issuance of suspect certificates as well as to audit the certificate logs themselves. The intent is that eventually clients would refuse to honor certificates that do not appear in a log, effectively forcing CAs to add all issued certificates to the logs. This document obsoletes RFC 6962. It also specifies a new TLS extension that is used to send various CT log artifacts. Logs are network services that implement the protocol operations for submissions and queries that are defined in this document. UTF-8, a transformation format of ISO 10646 ISO/IEC 10646-1 defines a large character set called the Universal Character Set (UCS) which encompasses most of the world's writing systems. The originally proposed encodings of the UCS, however, were not compatible with many current applications and protocols, and this has led to the development of UTF-8, the object of this memo. UTF-8 has the characteristic of preserving the full US-ASCII range, providing compatibility with file systems, parsers and other software that rely on US-ASCII values but are transparent to other values. This memo obsoletes and replaces RFC 2279. Automatic Certificate Management Environment (ACME) Public Key Infrastructure using X.509 (PKIX) certificates are used for a number of purposes, the most significant of which is the authentication of domain names. Thus, certification authorities (CAs) in the Web PKI are trusted to verify that an applicant for a certificate legitimately represents the domain name(s) in the certificate. As of this writing, this verification is done through a collection of ad hoc mechanisms. This document describes a protocol that a CA and an applicant can use to automate the process of verification and certificate issuance. The protocol also provides facilities for other certificate management functions, such as certificate revocation. Secure hash standard National Institute of Standards and Technology (U.S.) Internet X.509 Public Key Infrastructure Certificate and Certificate Revocation List (CRL) Profile 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] IEEE/Open Group Standard for Information Technology--Portable Operating System Interface (POSIX™) Base Specifications, Issue 8 IEEE X.509 Internet Public Key Infrastructure Online Certificate Status Protocol - OCSP This document specifies a protocol useful in determining the current status of a digital certificate without requiring Certificate Revocation Lists (CRLs). Additional mechanisms addressing PKIX operational requirements are specified in separate documents. This document obsoletes RFCs 2560 and 6277. It also updates RFC 5912. Unsigned X.509 Certificates This document defines a placeholder X.509 signature algorithm that may be used in contexts where the consumer of the certificate is not expected to verify the signature. As part of this, it updates RFC 5280. Web Linking This specification defines a model for the relationships between resources on the Web ("links") and the type of those relationships ("link relation types"). It also defines the serialisation of such links in HTTP headers with the Link header field. HTTP Semantics The Hypertext Transfer Protocol (HTTP) is a stateless application-level protocol for distributed, collaborative, hypertext information systems. This document describes the overall architecture of HTTP, establishes common terminology, and defines aspects of the protocol that are shared by all versions. In this definition are core protocol elements, extensibility mechanisms, and the "http" and "https" Uniform Resource Identifier (URI) schemes. This document updates RFC 3864 and obsoletes RFCs 2818, 7231, 7232, 7233, 7235, 7538, 7615, 7694, and portions of 7230. Informative References Chrome Certificate Transparency Policy Google Chrome Apple's Certificate Transparency policy Apple Component Updater Chromium Firefox Remote Settings Mozilla Let's Encrypt Stats Let's Encrypt Cloudflare Radar Certificate Transparency Cloudflare, Inc. Finding shared RSA factors in the Certificate Transparency logs University of Bergen Security in the Presence of Key Reuse: Context-Separable Interfaces and their Applications STH Discipline & Security Considerations Ballot 153 – Short-Lived Certificates CA/Browser Forum Ballot SC081v3: Introduce Schedule of Reducing Validity and Data Reuse Periods CA/Browser Forum How to distrust a CA without any certificate errors Private SCT Auditing, Revisited Merkle Tree Certificates With Tiled Transparency Logs C2SP Tiled Transparency Logs C2SP Transparency Log Witness Protocol C2SP Transparency Log Mirrors C2SP Transparency Log Cosignatures C2SP Accumulated Test Vectors Large Inclusion Proof Test Vectors IETF Large Consistency Proof Test Vectors IETF Certificate Transparency This document describes an experimental protocol for publicly logging the existence of Transport Layer Security (TLS) certificates as they are issued or observed, in a manner that allows anyone to audit certificate authority (CA) activity and notice the issuance of suspect certificates as well as to audit the certificate logs themselves. The intent is that eventually clients would refuse to honor certificates that do not appear in a log, effectively forcing CAs to add all issued certificates to the logs. Logs are network services that implement the protocol operations for submissions and queries that are defined in this document. Edwards-Curve Digital Signature Algorithm (EdDSA) This document describes elliptic curve signature scheme Edwards-curve Digital Signature Algorithm (EdDSA). The algorithm is instantiated with recommended parameters for the edwards25519 and edwards448 curves. An example implementation and test vectors are provided. Module-lattice-based digital signature standard National Institute of Standards and Technology (U.S.) Lightweight Directory Access Protocol (LDAP): String Representation of Distinguished Names The X.500 Directory uses distinguished names (DNs) as primary keys to entries in the directory. This document defines the string representation used in the Lightweight Directory Access Protocol (LDAP) to transfer distinguished names. The string representation is designed to give a clean representation of commonly used distinguished names, while being able to represent any distinguished name. [STANDARDS-TRACK] Privacy Considerations for Internet Protocols This document offers guidance for developing privacy considerations for inclusion in protocol specifications. It aims to make designers, implementers, and users of Internet protocols aware of privacy-related design choices. It suggests that whether any individual RFC warrants a specific privacy considerations section will depend on the document's content. Object Identifier Registry for the PKIX Working Group When the Public-Key Infrastructure using X.509 (PKIX) Working Group was chartered, an object identifier arc was allocated by IANA for use by that working group. This document describes the object identifiers that were assigned in that arc, returns control of that arc to IANA, and establishes IANA allocation policies for any future assignments within that arc.

ASN.1 Module

Merkle Tree Structure This non-normative section describes how the Merkle Tree structure relates to the binary representations of indices. It is included to help implementers understand the procedures described in .

Binary Representations Within a Merkle Tree whose size is a power of two, the binary representation of a leaf's index gives the path to that leaf. The leaf is a left child if the least-significant bit is unset and a right child if it is set. The next bit indicates the direction of the parent node, and so on. demonstrates this in a Merkle Tree of size 8:

An example Merkle Tree of size 8 The binary representation of 4 is 0b100. It is the left (0) child of [4, 6), which is the left (0) child of [4, 8), which is the right (1) child of [0, 8). Each level in the tree corresponds to a bit position and can be correspondingly numbered, with 0 indicating the least-significant bit and the leaf level, and so on. In this numbering, a node's level can be determined as follows: if the node is a root of subtree [start, end), the node's level is BIT_WIDTH(end - start - 1). Comparing two indices determines the relationship between two paths. The highest differing bit gives the level at which paths from root to leaf diverge. For example, the bit representations of 4 and 6 are 0b100 and 0b110, respectively. The highest differing bit is bit 1. Bits 2 and up are the same between the two indices. This indicates that the paths from the root to leaves 4 and 6 diverge when going from level 2 to level 1. This can be generalized to arbitrary-sized Merkle Trees. depicts a Merkle Tree of size 6:

An example Merkle Tree of size 6 When the size of a Merkle Tree is not a power of two, some levels on the rightmost edge of the tree are skipped. The rightmost edge is the path to the last element. The skipped levels can be seen in its binary representation. Here, the last element is 5, which has binary representation 0b101. When a bit is set, the corresponding node is a right child. When it is unset, the corresponding node is skipped. In a tree of the next power of two size, the skipped nodes in this path are where there would have been a right child, had there been enough elements to construct one. Without a right child, the hash operation is skipped and a skipped node has the same value as its singular child. depicts this for a tree of size 6.

An example Merkle Tree of size 6, viewed as a subset of a tree of size 8 Zero bits also indicate skipped nodes in paths that have not yet diverged from the rightmost edge (i.e. the path to the last element), when viewed from root to leaf. In the example, the binary representation of 4 is 0b100. While bit 0 and bit 1 are both unset, they manifest in the tree differently. Bit 0 indicates that 4 is a left child. However, at bit 1, 0b100 has not yet diverged from the last element, 0b101. That instead indicates a skipped node, not a left child.

Subtrees Given a list of elements and Merkle Tree over them, it is possible to construct a smaller Merkle Tree over any interval of elements. However, those smaller trees may not have the same structure as the original tree. shows a Merkle Tree of size 8, and a tree built over elements [1, 5). When [1, 5) is considered as an independent, 4-element sequence, it does not align with the portion of the overall tree that covers [1, 5). The two trees do not share any intermediate nodes. This prevents constructing subtree consistency proofs ().

An example misaligned tree The numerical constraints on start and end in restrict subtrees to ensure that they are properly aligned with the original tree as to permit subtree consistency proofs. A Merkle Tree built over [start, end) has size end - start, and is constructed as if start were the first element of the sequence at index zero. To be aligned, start must be the leftmost leaf of the lowest common ancestor of start and end - 1 in the original tree:

Numerically, this means the least significant BIT_WIDTH(end - start - 1) bits of start must be zero. Equivalently, start must be divisible by BIT_CEIL(end - start).
In the tree, this means subtrees are constructed by taking any node in the tree, setting start to the leftmost leaf under the node, and end to one past any other leaf under the node.

Though most nodes overlap, not every node of the subtree is necessarily in the larger Merkle Tree, as shown in . In general:

Subtrees whose sizes are a power of two are called full subtrees. A full subtree's root node will always be in the original tree.
Subtrees whose sizes are not a power of two are called partial subtrees. A partial subtree's root node will be in the original tree of size n, if and only if n = end. Otherwise, non-leaf nodes along the partial subtree's right edge will not be part of the original tree.

The difference between full and partial subtrees does not impact their usage, but they can help in understanding the proof constructions below.

Inclusion Proof Evaluation The procedure in builds up a subtree hash in r by starting from entry_hash and iteratively hashing elements of inclusion_proof on the left or right. That means this procedure, when successful, must return some hash that contains entry_hash. Treating [start, end) as a Merkle Tree of size end - start, the procedure hashes based on the path to index. Within this smaller Merkle Tree, it has index fn = index - start (first number), and the last element has index sn = end - start - 1 (second number). Step 4 iterates through inclusion_proof and the paths to fn and sn in parallel. As the procedure right-shifts fn and sn and looks at the least-significant bit, it moves up the two paths, toward the root. When sn is zero, the procedure has reached the top of the tree. The procedure checks that the two iterations complete together. Iterating from level 0 up, fn and sn will initially be different. While they are different, step 4.2 hashes on the left or right based on the binary representation, as discussed in . Once fn = sn, the remainder of the path is on the right edge. At that point, the condition in step 4.2 is always true. It only incorporates proof entries on the left, once per set bit. Unset bits are skipped. Inclusion proofs can also be evaluated by considering these two stages separately. The first stage consumes l1 = BIT_WIDTH(fn XOR sn) proof entries. The second stage consumes l2 = POPCOUNT(fn >> l1) proof entries. A valid inclusion proof must then have l1 + l2 entries. The first l1 entries are hashed based on fn's least significant bits, and the remaining l2 entries are hashed on the left.

Consistency Proof Structure A subtree consistency proof for [start, end) and the tree of n elements is similar to an inclusion proof for element end - 1. If one starts from end - 1's hash, incorporating the whole inclusion proof should reconstruct root_hash and incorporating a subset of the inclusion proof should reconstruct node_hash. Thus end - 1's hash and this inclusion proof can prove consistency. A subtree consistency proof in this document applies two optimizations over this construction:

Instead of starting at level 0 with end - 1, the proof can start at a higher level. Any ancestor of end - 1 shared by both the subtree and the overall tree is a valid starting node to reconstruct node_hash and root_hash. Use the highest level with a common ancestor. This truncates the inclusion proof.
If this starting node is the entire subtree, omit its hash from the consistency proof. The verifier is assumed to already know node_hash.

A Merkle consistency proof, defined in , applies these same optimizations. depicts a subtree consistency proof between the subtree [0, 6) and the Merkle Tree of size 8. The consistency proof begins at level 1, or node [4, 6). The inclusion proof portion is similarly truncated to start at level 1: [6, 8) and [0, 4). If the consistency proof began at level 0, the starting node would be leaf 5, and the consistency proof would additionally include leaf 4.

A subtree consistency proof that starts at level 1 instead of level 0 Note that the truncated inclusion proof may include nodes from lower levels, if the corresponding level was skipped on the right edge. depicts a subtree consistency proof between the subtree [0, 6) and the Merkle Tree of size 7. As above, the starting node is [4, 6) at level 1. The inclusion proof portion includes leaf 6 at level 0. This is because leaf 6 is taking the place of its skipped parent at level 1. (A skipped node can be thought of as a duplicate of its singular child.)

The interaction between inclusion proof truncation and skipped levels

Consistency Proof Verification The procedure in is structured similarly to inclusion proof evaluation (). It iteratively builds two hashes, fr (first root) and sr (second root), which are expected to equal node_hash and root_hash, respectively. Everything hashed into fr is also hashed into sr, so success demonstrates that root_hash contains node_hash. Step 2 initializes fn (first number), sn (second number), and tn (third number) to follow, respectively, the paths to start, end - 1 (the last element of the subtree), and n - 1 (the last element of the tree). Steps 3 and 4 then skip to the starting node, described in . The starting node may be:

The entire subtree [start, end) if the subtree root is in the tree. This will occur if end is n (step 3), or if [start, end) is a full subtree (exiting step 4 because fn is sn).
Otherwise, the highest full subtree along the right edge of [start, end). This corresponds to the process exiting step 4 because LSB(sn) is not set.

Steps 5 and 6 initialize the hashes fr and sr:

In the first case above, fn will equal sn after truncation. Step 5 will then initialize the hashes to node_hash because the consistency proof does not need to include the starting node.
In the second case above, fn is less than sn. Step 6 will then initialize the hashes to the first value in the consistency proof.

Step 7 incorporates the remainder of the consistency proof into fr and sr:

All hashes are incorporated into sr, with hashing on the left or right determined the same as in inclusion proof evaluation.
A subset of the hashes is incorporated into fr. It skips any hash on the right because those contain elements greater than end - 1. It also stops incorporating when fn and sn have converged.

This reconstructs the hashes of the subtree and original tree, which are then compared to expected values in step 8. In the case when fn is sn in step 5, the condition in step 7.2.1 is always false, and fr is always equal to node_hash in step 8. In this case, steps 6 through 8 are equivalent to verifying an inclusion proof for the truncated subtree [fn, sn + 1) and truncated tree tn + 1.

Test Vectors

Accumulated Subtree Test Vectors The following are "accumulated" test vectors for the various subtree algorithms defined in . They are hash values of the outputs of all possible inputs for each algorithm, for trees of sizes up to 130. They can be used to verify that an implementation matches the specification, without having to include a large number of individual test vectors. For all the test vectors, a tree D_n of size n is constructed with leaf values d[0] = 0x00, d[1] = 0x01, .... The hash function used is SHA-256. The hash values are encoded in hexadecimal.

Subtree Hashes For each value of end from 0 to 130, and each value of start from 0 to end, if [start, end) is a valid subtree, add to the rolling hash the ASCII string [START, END) HASH followed by a newline (U+000A), where START and END are the decimal representations of start and end, respectively, and HASH is the hexadecimal encoding of MTH(D[start:end]), according to . The final hash value is In Python, this can be expressed as: This test exercises both an implementation's subtree hash calculation as well as the subtree validity check. In a CA, both of these operations apply. In a relying party, only the subtree validity check applies. A relying party implementation SHOULD implement a Merkle Tree in testing logic so the above will exercise the subtree validity check.

Subtree Inclusion Proofs For each value of end from 0 to 130, and each value of start from 0 to end, if [start, end) is a valid subtree, for each value of index from start to end - 1, add to the rolling hash the ASCII string INDEX [START, END), then, for each hash in the inclusion proof () for d[index] in the subtree [start, end), a space (U+0020) followed by the hexadecimal encoding of that hash, and finally a newline (U+000A), where INDEX is the decimal representation of index, and START and END are the decimal representations of start and end, respectively. The final hash value is In Python, this can be expressed as: This test exercises constructing subtree inclusion proofs, as computed by a CA. A relying party instead evaluates untrusted subtree inclusion proofs. This test can be used to exercise this logic as follows:

If not available, implement a Merkle Tree in testing logic to compute subtree hashes and subtree inclusion proofs. This logic can be validated by the above test and .
For each subtree inclusion proof in the above test, evaluate the inclusion proof and assert the resulting hash matches the corresponding subtree hash.
For each non-empty subtree inclusion proof in the above test, truncate the proof by both one byte and a full hash. Assert that evaluating each proof fails.
For each subtree inclusion proof in the above test, extend the proof by both one byte and an arbitrary full hash. Assert that evaluating each proof fails.

Subtree Consistency Proofs For each value of n from 0 to 130, and each value of end from 0 to n, and each value of start from 0 to end, if [start, end) is a valid subtree, add to the rolling hash the ASCII string [START, END) N, then, for each hash in the consistency proof () for the subtree [start, end) and tree of size n, a space (U+0020) followed by the hexadecimal encoding of that hash, and finally a newline (U+000A), where START and END are the decimal representations of start and end, respectively, and N is the decimal representation of n. The final hash value is In Python, this can be expressed as: This test exercises constructing subtree consistency proofs. Other parties will check these proofs (e.g. and ). This test can be used to exercise this logic as follows:

If not available, implement a Merkle Tree in testing logic to compute subtree hashes and subtree consistency proofs. This logic can be validated by the above test and .
For each subtree consistency proof in the above test, check the proof and assert it succeeds.
For each non-empty subtree consistency proof in the above test, truncate the proof by both one byte and a full hash. Assert that checking each proof fails.
For each subtree consistency proof in the above test, extend the proof by both one byte and an arbitrary full hash. Assert that checking each proof fails.
For each subtree consistency proof in the above test, flip a bit in the input subtree hash. Assert that checking each proof fails.
For each subtree consistency proof in the above test with a non-empty subtree, flip a bit in the input tree hash. Assert that checking each proof fails.

Efficient Covering Subtrees For each value of end from 0 to 130, and each value of start from 0 to end, add to the rolling hash the ASCII string [LEFT_START, LEFT_END) [RIGHT_START, RIGHT_END) followed by a newline (U+000A), where LEFT_START, LEFT_END, RIGHT_START, and RIGHT_END are the decimal representations of the start and end of the left and right subtrees, respectively, that efficiently cover () [start, end). The final hash value is In Python, this can be expressed as:

Large Subtree Test Vectors exhaustively tests subtree algorithms for trees of size up to 130. An implementation may also encounter overflow conditions with very large trees. This is primarily a concern for a relying party, which must act on untrusted tree sizes. The largest possible Merkle Tree in this protocol is 2⁴⁸-1. In particular, the MTCProof structure in cannot express larger values. These sizes will fit comfortably in 64-bit integers, whether signed or unsigned. To exercise more general implementations, this section includes test vectors for trees bounded by 2⁴⁸-1, 2⁶³-1, and 2⁶⁴-1. Implementations MAY skip tests above 2⁴⁸-1 if they do not support such trees.

Subtree Validity The following are valid subtrees ():

start is 0 and end is 2⁴⁷ + 1
start is 0 and end is 2⁴⁸ - 1
start is 0 and end is 2⁶² + 1
start is 0 and end is 2⁶³ - 1
start is 0 and end is 2⁶³ + 1
start is 0 and end is 2⁶⁴ - 1

The following are not valid subtrees:

start is 2⁴⁶ and end is 2⁴⁷ + 1
start is 2⁴⁶ and end is 2⁴⁸ - 1
start is 2⁶¹ and end is 2⁶² + 1
start is 2⁶¹ and end is 2⁶³ - 1
start is 2⁶² and end is 2⁶³ + 1
start is 2⁶² and end is 2⁶⁴ - 1

Subtree Inclusion Proofs contains additional test vectors for subtree inclusion proofs ().

Subtree Consistency Proofs contains additional test vectors for subtree consistency proofs ().

Efficient Covering Subtrees This section contains sample inputs and outputs for the procedure in . Given range [0x0, 0x800000000000), the subtrees are:

[0x0, 0x400000000000)
[0x400000000000, 0x800000000000)

Given range [0x500000000000, 0xd00000000000), the subtrees are:

[0x400000000000, 0x800000000000)
[0x800000000000, 0xd00000000000)

Given range [0x7fffffffffff, 0x800000000001), the subtrees are:

[0x7fffffffffff, 0x800000000000)
[0x800000000000, 0x800000000001)

Given range [0xfffffffffffe, 0xffffffffffff), the subtrees are:

[0xfffffffffffe, 0xffffffffffff)
[0xffffffffffff, 0xffffffffffff)

Given range [0xffffffffffff, 0xffffffffffff), the subtrees are:

[0xffffffffffff, 0xffffffffffff)
[0xffffffffffff, 0xffffffffffff)

Given range [0x0, 0x4000000000000000), the subtrees are:

[0x0, 0x2000000000000000)
[0x2000000000000000, 0x4000000000000000)

Given range [0x2800000000000000, 0x6800000000000000), the subtrees are:

[0x2000000000000000, 0x4000000000000000)
[0x4000000000000000, 0x6800000000000000)

Given range [0x3fffffffffffffff, 0x4000000000000001), the subtrees are:

[0x3fffffffffffffff, 0x4000000000000000)
[0x4000000000000000, 0x4000000000000001)

Given range [0x7ffffffffffffffe, 0x7fffffffffffffff), the subtrees are:

[0x7ffffffffffffffe, 0x7fffffffffffffff)
[0x7fffffffffffffff, 0x7fffffffffffffff)

Given range [0x7fffffffffffffff, 0x7fffffffffffffff), the subtrees are:

[0x7fffffffffffffff, 0x7fffffffffffffff)
[0x7fffffffffffffff, 0x7fffffffffffffff)

Given range [0x0, 0x8000000000000000), the subtrees are:

[0x0, 0x4000000000000000)
[0x4000000000000000, 0x8000000000000000)

Given range [0x5000000000000000, 0xd000000000000000), the subtrees are:

[0x4000000000000000, 0x8000000000000000)
[0x8000000000000000, 0xd000000000000000)

Given range [0x7fffffffffffffff, 0x8000000000000001), the subtrees are:

[0x7fffffffffffffff, 0x8000000000000000)
[0x8000000000000000, 0x8000000000000001)

Given range [0xfffffffffffffffe, 0xffffffffffffffff), the subtrees are:

[0xfffffffffffffffe, 0xffffffffffffffff)
[0xffffffffffffffff, 0xffffffffffffffff)

Given range [0xffffffffffffffff, 0xffffffffffffffff), the subtrees are:

[0xffffffffffffffff, 0xffffffffffffffff)
[0xffffffffffffffff, 0xffffffffffffffff)

Acknowledgements This document stands on the shoulders of giants and builds upon decades of work in TLS authentication, X.509, and Certificate Transparency. The authors would like to thank all those who have contributed over the history of these protocols. The authors additionally thank Bob Beck, Corey Bonnell, Ryan Dickson, Aaron Gable, Nick Harper, Jacob Hoffman-Andrews, Russ Housley, Dennis Jackson, Ilari Liusvaara, Sanketh Menda, Matt Mueller, Mike Ounsworth, Chris Patton, Michael Richardson, Ryan Sleevi, Emily Stark, and Rob Stradling for many valuable discussions and insights which led to this document, as well as feedback and contributions to the document itself. We wish to thank Mia Celeste in particular, whose implementation of an earlier draft revealed several pitfalls. The idea to mint tree heads infrequently was originally described by Richard Barnes in . The size optimization in Merkle Tree Certificates is an application of this idea to the certificate itself.

Change log

RFC Editor's Note: Please remove this section prior to publication of a final version of this document.

Since draft-davidben-tls-merkle-tree-certs-00

Simplify hashing by removing the internal padding to align with block size. #72
Avoid the temptation of floating points. #66
Require lifetime to be a multiple of batch_duration. #65
Rename window to validity window. #21
Split Assertion into Assertion and AbridgedAssertion. The latter is used in the Merkle Tree and HTTP interface. It replaces subject_info by a hash, to save space by not serving large post-quantum public keys. The original Assertion is used everywhere else, including BikeshedCertificate. #6
Add proper context to every node in the Merkle Tree. #32
Clarify we use a single CertificateEntry. #11
Clarify we use POSIX time. #1
Elaborate on CA public key and signature format. #27
Miscellaneous changes.

Since draft-davidben-tls-merkle-tree-certs-01

Minor editorial changes

Since draft-davidben-tls-merkle-tree-certs-02

Replace the negotiation mechanism with TLS Trust Anchor Identifiers.

Since draft-davidben-tls-merkle-tree-certs-03

Switch terminology from "subscriber" to "authenticating party".
Use <1..2^24-1> encoding for all certificate types in the CertificateEntry TLS message
Clarify discussion and roles in transparency ecosystem
Update references

Since draft-davidben-tls-merkle-tree-certs-04 Substantially reworked the design. The old design was essentially the landmark checkpoint and CA-built logs ideas, but targeting only the optimized and slow issuance path, and with a more bespoke tree structure: In both draft-04 and draft-05, a CA looks like today's CAs except that they run some software to publish what they issue and sign tree heads to certify certificates in bulk. In draft-04, the CA software publishes certificates in a bunch of independent Merkle Trees. This is very easy to do as a collection of highly cacheable, immutable static files because each tree is constructed independently, and never appended to after being built. In draft-05, the certificates are published in a single Merkle Tree. The interface allows such trees to also use highly cacheable, immutable static files. In draft-04, there only are hourly tree heads. Clients are provisioned with tree heads ahead of time so we can make small, inclusion-proof-only certificates. In draft-05, the ecosystem must coordinate on defining "landmark" checkpoints. Clients are provisioned with subtrees describing landmark checkpoints ahead of time so we can make small, inclusion-proof-only certificates. In draft-04, each tree head is independent. In draft-05, each landmark checkpoint contains all the previous checkpoints. In draft-04, the independent tree heads were easily prunable. In draft-05, we define how to prune a Merkle Tree. In draft-04, there is no fast issuance mode. In draft-05, frequent, non-landmark checkpoints can be combined with inclusion proofs and witness signatures for fast issuance. This is essentially an STH and inclusion proof in CT.

Since draft-davidben-tls-merkle-tree-certs-05

Add some discussion on malleability
Discuss the monitoring impacts of the responsibility shift from CA with log quorum to CA+log with mirror quorum
Sketch out a more concrete initial ACME extension

Since draft-davidben-tls-merkle-tree-certs-06

Fix mistyped reference
Removed now unnecessary placeholder text
First draft at IANA registration and ASN.1 module
Added a prose version of the procedure to select subtrees
Rename 'landmarks checkpoint' to 'landmarks'
Clarify and fix an off-by-one error in recommended landmark allocation scheme
Add some diagrams to the Overview section

Since draft-davidben-tls-merkle-tree-certs-07

Clarify landmark zero
Clarify signature verification process
Improve subtree consistency proof verification algorithm
Add an appendix that explains the Merkle Tree proof procedures

Since draft-davidben-tls-merkle-tree-certs-08

Improvements to malleability discussion
Improvements to subtree definition
Improvements to trust_anchors integration

Since draft-davidben-tls-merkle-tree-certs-09

Editorial fixes
Set a more accurate intended status
Fixes to ASN.1 module
Make log entry more friendly to single-pass verification

Since draft-davidben-tls-merkle-tree-certs-10

Adopted by working group

Since draft-ietf-plants-merkle-tree-certs-00

Address editorial comments from WG adoption call

Since draft-ietf-plants-merkle-tree-certs-01

Renamed full certificate to standalone certificate, signatureless certificate to landmark certificate.
Included subject public key algorithm in log entries

Since draft-ietf-plants-merkle-tree-certs-02

Renamed landmark certificate to landmark-relative certificate
Relaxed restrictions on null_entry
Clarify that CRLs and OCSPs apply to MTCs unmodified

Since draft-ietf-plants-merkle-tree-certs-03

Use a tlog-compatible signature scheme for ease of deployment
Define a CA certificate representation
Remove the one-to-many relationship between MTC CAs and CA cosigners
Discuss domain separation for signatures
Recommend a maximum log entry size for tlog compatibility
Prescribe landmark OID allocation
Update TLS integration now that trust anchor IDs extension has been moved to the base draft
A single CA now operates a series of issuance logs, instead of a one-to-one correspondence
Group components of a CA into a CA-specific section that enumerates the parts of a CA
Canonicalize the order of cosignatures in MTCProofs
Remove sketch of tlog subtree signer API in favor of https://github.com/C2SP/C2SP/pull/245 in
Add an extensions block to log entries

Since draft-ietf-plants-merkle-tree-certs-04

Fix some mistakes in the single-pass signature verification algorithm
Editorial fixes
Discuss the implications of subordinate CAs in Security Considerations
Added subtree test vector appendix
Define a CA's current issuance log and rules around that
Switch the ACME construction to a new link relation and change the HTTP status code
Add a maxSerial field to the CA format

Since draft-ietf-plants-merkle-tree-certs-05

Use 24-bit length prefix for MTCProof subtree signatures.
Renamed MerkleTreeCertEntry, etc., structures to MTCLogEntry to be consistent with MTCProof, shorter, and help disambiguate the many English meanings of "entry".
Fixed one of the accumulated test vectors to better reflect one of the edge cases in subtree covering.
Make empty subtrees valid, so the subtree covering function always returns two subtrees.
Add an informative reference to the MTC-TLOG profile (c2sp.org/mtc-tlog) and mention it where tile-based logs are discussed.
Clarify that a landmark consists of both a number and a tree size, and that a landmark's subtrees share its landmark number.
Give an exact procedure for selecting the landmark and covering subtree when constructing a landmark-relative certificate.
Prune the pruning discussion. It's really a property of the log serving protocol and is better described in and .
Fix the maximum log index to account for also end being 48-bit.
Discuss a potential overflow in the valid subtree definition.
Describe how a party holding a standalone certificate can construct the corresponding landmark-relative certificate itself.
Added test vectors for subtree algorithms in larger trees.
Align the experimental OID with the final one in the X.509 name construction in using RELATIVE-OID directly.
Lifted the tree hash into the MTC CA extension OID, so it can capture new tree constructions more generally.
Update for draft-ietf-tls-trust-anchor-ids-05, and spell out certificate configuration explicitly.
Define active landmarks around landmark expiry and put the expiry time in the landmark format.
Fix an interaction between landmark-relative certs and TLS certificate_authorities.