| Public Notary Transparency Working Group | B. Laurie |
| Internet-Draft | A. Langley |
| Intended status: Standards Track | E. Kasper |
| Expires: May 23, 2016 | E. Messeri |
| R. Stradling | |
| Comodo | |
| November 20, 2015 |
Certificate Transparency
draft-ietf-trans-rfc6962-bis-11
This document describes a 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 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.
Logs are network services that implement the protocol operations for submissions and queries that are defined in this document.
This Internet-Draft is submitted in full conformance with the provisions of BCP 78 and BCP 79.
Internet-Drafts are working documents of the Internet Engineering Task Force (IETF). Note that other groups may also distribute working documents as Internet-Drafts. The list of current Internet-Drafts is at http://datatracker.ietf.org/drafts/current/.
Internet-Drafts are draft documents valid for a maximum of six months and may be updated, replaced, or obsoleted by other documents at any time. It is inappropriate to use Internet-Drafts as reference material or to cite them other than as "work in progress."
This Internet-Draft will expire on May 23, 2016.
Copyright (c) 2015 IETF Trust and the persons identified as the document authors. All rights reserved.
This document is subject to BCP 78 and the IETF Trust's Legal Provisions Relating to IETF Documents (http://trustee.ietf.org/license-info) in effect on the date of publication of this document. Please review these documents carefully, as they describe your rights and restrictions with respect to this document. Code Components extracted from this document must include Simplified BSD License text as described in Section 4.e of the Trust Legal Provisions and are provided without warranty as described in the Simplified BSD License.
Certificate transparency aims to mitigate the problem of misissued certificates by providing publicly auditable, append-only, untrusted logs of all issued certificates. The logs are publicly auditable so that it is possible for anyone to verify the correctness of each log and to monitor when new certificates are added to it. The logs do not themselves prevent misissue, but they ensure that interested parties (particularly those named in certificates) can detect such misissuance. Note that this is a general mechanism, but in this document, we only describe its use for public TLS server certificates issued by public certification authorities (CAs).
Each log consists of certificate chains, which can be submitted by anyone. It is expected that public CAs will contribute all their newly issued certificates to one or more logs, however certificate holders can also contribute their own certificate chains, as can third parties. In order to avoid logs being rendered useless by submitting large numbers of spurious certificates, it is required that each chain is rooted in a CA certificate accepted by the log. When a chain is submitted to a log, a signed timestamp is returned, which can later be used to provide evidence to TLS clients that the chain has been submitted. TLS clients can thus require that all certificates they accept as valid are accompanied by signed timestamps.
Those who are concerned about misissue can monitor the logs, asking them regularly for all new entries, and can thus check whether domains they are responsible for have had certificates issued that they did not expect. What they do with this information, particularly when they find that a misissuance has happened, is beyond the scope of this document, but broadly speaking, they can invoke existing business mechanisms for dealing with misissued certificates, such as working with the CA to get the certificate revoked, or with maintainers of trust anchor lists to get the CA removed. Of course, anyone who wants can monitor the logs and, if they believe a certificate is incorrectly issued, take action as they see fit.
Similarly, those who have seen signed timestamps from a particular log can later demand a proof of inclusion from that log. If the log is unable to provide this (or, indeed, if the corresponding certificate is absent from monitors' copies of that log), that is evidence of the incorrect operation of the log. The checking operation is asynchronous to allow TLS connections to proceed without delay, despite network connectivity issues and the vagaries of firewalls.
The append-only property of each log is technically achieved using Merkle Trees, which can be used to show that any particular instance of the log is a superset of any particular previous instance. Likewise, Merkle Trees avoid the need to blindly trust logs: if a log attempts to show different things to different people, this can be efficiently detected by comparing tree roots and consistency proofs. Similarly, other misbehaviors of any log (e.g., issuing signed timestamps for certificates they then don't log) can be efficiently detected and proved to the world at large.
The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this document are to be interpreted as described in RFC 2119 [RFC2119].
Data structures are defined according to the conventions laid out in Section 4 of [RFC5246].
Logs use a binary Merkle Hash Tree for efficient auditing. The hashing algorithm used by each log is expected to be specified as part of the metadata relating to that log. We have established a registry of acceptable algorithms, see Section 11.2. The hashing algorithm in use is referred to as HASH throughout this document and the size of its output in bytes as HASH_SIZE. The input to the Merkle Tree Hash is a list of data entries; these entries will be hashed to form the leaves of the Merkle Hash Tree. The output is a single HASH_SIZE Merkle Tree Hash. Given an ordered list of n inputs, D[n] = {d(0), d(1), ..., d(n-1)}, the Merkle Tree Hash (MTH) is thus defined as follows:
The hash of an empty list is the hash of an empty string:
MTH({}) = HASH().
The hash of a list with one entry (also known as a leaf hash) is:
MTH({d(0)}) = HASH(0x00 || d(0)).
For n > 1, let k be the largest power of two smaller than n (i.e., k < n <= 2k). The Merkle Tree Hash of an n-element list D[n] is then defined recursively as
MTH(D[n]) = HASH(0x01 || MTH(D[0:k]) || MTH(D[k:n])),
where || is concatenation and D[k1:k2] denotes the list {d(k1), d(k1+1),..., d(k2-1)} of length (k2 - k1). (Note that the hash calculations for leaves and nodes differ. This domain separation is required to give second preimage resistance.)
Note that we do not require the length of the input list to be a power of two. The resulting Merkle Tree may thus not be balanced; however, its shape is uniquely determined by the number of leaves. (Note: This Merkle Tree is essentially the same as the history tree [CrosbyWallach] proposal, except our definition handles non-full trees differently.)
A Merkle inclusion proof for a leaf in a Merkle Hash Tree is the shortest list of additional nodes in the Merkle Tree required to compute the Merkle Tree Hash for that tree. Each node in the tree is either a leaf node or is computed from the two nodes immediately below it (i.e., towards the leaves). At each step up the tree (towards the root), a node from the inclusion proof is combined with the node computed so far. In other words, the inclusion proof consists of the list of missing nodes required to compute the nodes leading from a leaf to the root of the tree. If the root computed from the inclusion proof matches the true root, then the inclusion proof proves that the leaf exists in the tree.
Given an ordered list of n inputs to the tree, D[n] = {d(0), ..., d(n-1)}, the Merkle inclusion proof PATH(m, D[n]) for the (m+1)th input d(m), 0 <= m < n, is defined as follows:
The proof for the single leaf in a tree with a one-element input list D[1] = {d(0)} is empty:
PATH(0, {d(0)}) = {}
For n > 1, let k be the largest power of two smaller than n. The proof for the (m+1)th element d(m) in a list of n > m elements is then defined recursively as
PATH(m, D[n]) = PATH(m, D[0:k]) : MTH(D[k:n]) for m < k; and
PATH(m, D[n]) = PATH(m - k, D[k:n]) : MTH(D[0:k]) for m >= k,
where : is concatenation of lists and D[k1:k2] denotes the length (k2 - k1) list {d(k1), d(k1+1),..., d(k2-1)} as before.
Merkle consistency proofs prove the append-only property of the tree. A Merkle consistency proof for a Merkle Tree Hash MTH(D[n]) and a previously advertised hash MTH(D[0:m]) of the first m leaves, m <= n, is the list of nodes in the Merkle Tree required to verify that the first m inputs D[0:m] are equal in both trees. Thus, a consistency proof must contain a set of intermediate nodes (i.e., commitments to inputs) sufficient to verify MTH(D[n]), such that (a subset of) the same nodes can be used to verify MTH(D[0:m]). We define an algorithm that outputs the (unique) minimal consistency proof.
Given an ordered list of n inputs to the tree, D[n] = {d(0), ..., d(n-1)}, the Merkle consistency proof PROOF(m, D[n]) for a previous Merkle Tree Hash MTH(D[0:m]), 0 < m < n, is defined as:
PROOF(m, D[n]) = SUBPROOF(m, D[n], true)
The subproof for m = n is empty if m is the value for which PROOF was originally requested (meaning that the subtree Merkle Tree Hash MTH(D[0:m]) is known):
SUBPROOF(m, D[m], true) = {}
The subproof for m = n is the Merkle Tree Hash committing inputs D[0:m]; otherwise:
SUBPROOF(m, D[m], false) = {MTH(D[m])}
For m < n, let k be the largest power of two smaller than n. The subproof is then defined recursively.
If m <= k, the right subtree entries D[k:n] only exist in the current tree. We prove that the left subtree entries D[0:k] are consistent and add a commitment to D[k:n]:
SUBPROOF(m, D[n], b) = SUBPROOF(m, D[0:k], b) : MTH(D[k:n])
If m > k, the left subtree entries D[0:k] are identical in both trees. We prove that the right subtree entries D[k:n] are consistent and add a commitment to D[0:k].
SUBPROOF(m, D[n], b) = SUBPROOF(m - k, D[k:n], false) : MTH(D[0:k])
Here, : is a concatenation of lists, and D[k1:k2] denotes the length (k2 - k1) list {d(k1), d(k1+1),..., d(k2-1)} as before.
The number of nodes in the resulting proof is bounded above by ceil(log2(n)) + 1.
The binary Merkle Tree with 7 leaves:
hash
/ \
/ \
/ \
/ \
/ \
k l
/ \ / \
/ \ / \
/ \ / \
g h i j
/ \ / \ / \ |
a b c d e f d6
| | | | | |
d0 d1 d2 d3 d4 d5
The inclusion proof for d0 is [b, h, l].
The inclusion proof for d3 is [c, g, l].
The inclusion proof for d4 is [f, j, k].
The inclusion proof for d6 is [i, k].
The same tree, built incrementally in four steps:
hash0 hash1=k
/ \ / \
/ \ / \
/ \ / \
g c g h
/ \ | / \ / \
a b d2 a b c d
| | | | | |
d0 d1 d0 d1 d2 d3
hash2 hash
/ \ / \
/ \ / \
/ \ / \
/ \ / \
/ \ / \
k i k l
/ \ / \ / \ / \
/ \ e f / \ / \
/ \ | | / \ / \
g h d4 d5 g h i j
/ \ / \ / \ / \ / \ |
a b c d a b c d e f d6
| | | | | | | | | |
d0 d1 d2 d3 d0 d1 d2 d3 d4 d5
The consistency proof between hash0 and hash is PROOF(3, D[7]) = [c, d, g, l]. c, g are used to verify hash0, and d, l are additionally used to show hash is consistent with hash0.
The consistency proof between hash1 and hash is PROOF(4, D[7]) = [l]. hash can be verified using hash1=k and l.
The consistency proof between hash2 and hash is PROOF(6, D[7]) = [i, j, k]. k, i are used to verify hash2, and j is additionally used to show hash is consistent with hash2.
Various data structures are signed. A log MUST use either deterministic ECDSA [RFC6979] using the NIST P-256 curve (Section D.1.2.3 of the Digital Signature Standard [DSS]) and HMAC-SHA256 or RSA signatures (RSASSA-PKCS1-v1_5 with SHA-256, Section 8.2 of [RFC3447]) using a key of at least 2048 bits.
Submitters submit certificates or precertificates to logs for public auditing, as described below. In order to enable attribution of each logged certificate or precertificate to its issuer, each submission MUST be accompanied by all additional certificates required to verify the chain up to an accepted root certificate. The root certificate itself MAY be omitted from the submission.
If a log accepts a submission, it will return a Signed Certificate Timestamp (SCT) (see Section 5.6). The submitter SHOULD validate the returned SCT as described in Section 9.2 if they understand its format and they intend to use it directly in a TLS handshake or to construct a certificate.
Anyone can submit a certificate [add-chain] to a log. Since certificates may not be accepted by TLS clients unless logged, it is expected that certificate owners or their CAs will usually submit them.
Alternatively, (root as well as intermediate) CAs may preannounce a certificate prior to issuance by submitting a precertificate [add-pre-chain] that the log can use to create an entry that will be valid against the issued certificate. The CA MAY incorporate the returned SCT in the issued certificate.
A precertificate is a CMS [RFC5652] signed-data object that conforms to the following requirements:
Some regard some DNS domain name labels within their registered domain space as private and security sensitive. Even though these domains are often only accessible within the domain owner's private network, it's common for them to be secured using publicly trusted TLS server certificates. We define a mechanism to allow these private labels to not appear in public logs.
A certificate containing a DNS-ID [RFC6125] of *.example.com could be used to secure the domain topsecret.example.com, without revealing the string topsecret publicly.
Since TLS clients only match the wildcard character to the complete leftmost label of the DNS domain name (see Section 6.4.3 of [RFC6125]), this approach would not work for a DNS-ID such as top.secret.example.com. Also, wildcard certificates are prohibited in some cases, such as Extended Validation Certificates [EVSSLGuidelines].
When creating a precertificate, the CA MAY substitute one or more labels in each DNS-ID with a corresponding number of ? labels. Every label to the left of a ? label MUST also be redacted. For example, if a certificate contains a DNS-ID of top.secret.example.com, then the corresponding precertificate could contain ?.?.example.com instead, but not top.?.example.com instead.
Wildcard * labels MUST NOT be redacted. However, if the complete leftmost label of a DNS-ID is *, it is considered redacted for the purposes of determining if the label to the right may be redacted. For example, if a certificate contains a DNS-ID of *.top.secret.example.com, then the corresponding precertificate could contain *.?.?.example.com instead, but not ?.?.?.example.com instead.
When a precertificate contains one or more redacted labels, a non-critical extension (OID 1.3.6.1.4.1.11129.2.4.6, whose extnValue OCTET STRING contains an ASN.1 SEQUENCE OF INTEGERs) MUST be added to the corresponding certificate: the first INTEGER indicates the total number of redacted labels and wildcard * labels in the precertificate's first DNS-ID; the second INTEGER does the same for the precertificate's second DNS-ID; etc. There MUST NOT be more INTEGERs than there are DNS-IDs. If there are fewer INTEGERs than there are DNS-IDs, the shortfall is made up by implicitly repeating the last INTEGER. Each INTEGER MUST have a value of zero or more. The purpose of this extension is to enable TLS clients to accurately reconstruct the TBSCertificate component of the precertificate from the certificate without having to perform any guesswork.
When a precertificate contains that extension and contains a CN-ID [RFC6125], the CN-ID MUST match the first DNS-ID and have the same labels redacted. TLS clients will use the first entry in the SEQUENCE OF INTEGERs to reconstruct both the first DNS-ID and the CN-ID.
An intermediate CA certificate or intermediate CA precertificate that contains the critical or non-critical Name Constraints [RFC5280] extension MAY be logged in place of end-entity certificates issued by that intermediate CA, as long as all of the following conditions are met:
Below is an example Name Constraints extension that meets these conditions:
SEQUENCE {
OBJECT IDENTIFIER '2 5 29 30'
OCTET STRING, encapsulates {
SEQUENCE {
[0] {
SEQUENCE {
[2] 'example.com'
}
}
[1] {
SEQUENCE {
[7] 00 00 00 00 00 00 00 00
}
SEQUENCE {
[7]
00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
}
}
}
}
}
A log is a single, append-only Merkle Tree of submitted certificate and precertificate entries.
When it receives a valid submission, the log MUST return an SCT that corresponds to the submitted certificate or precertificate. If the log has previously seen this valid submission, it MAY return the same SCT as it returned before. (Note that if a certificate was previously logged as a precertificate, then the precertificate's SCT of type precert_sct would not be appropriate; instead, a fresh SCT of type x509_sct should be generated).
An SCT is the log's promise to incorporate the submitted entry in its Merkle Tree no later than a fixed amount of time, known as the Maximum Merge Delay (MMD), after the issuance of the SCT. Periodically, the log MUST append all its new entries to its Merkle Tree and sign the root of the tree. This provides auditable evidence that the log kept all its promises.
Log operators MUST NOT impose any conditions on retrieving or sharing data from the log.
Logs MUST verify that each submitted certificate or precertificate has a valid signature chain to an accepted root certificate, using the chain of intermediate CA certificates provided by the submitter. Logs MUST accept certificates and precertificates that are fully valid according to RFC 5280 [RFC5280] verification rules and are submitted with such a chain. Logs MAY accept certificates and precertificates that have expired, are not yet valid, have been revoked, or are otherwise not fully valid according to RFC 5280 verification rules in order to accommodate quirks of CA certificate-issuing software. However, logs MUST reject submissions without a valid signature chain to an accepted root certificate. Logs MUST also reject precertificates that do not conform to the requirements in Section 3.2.
Logs SHOULD limit the length of chain they will accept. The maximum chain length is specified in the log's metadata.
The log SHALL allow retrieval of its list of accepted root certificates (see Section 6.8). This list might usefully be the union of root certificates trusted by major browser vendors.
If a submission is accepted and an SCT issued, the accepting log MUST store the entire chain used for verification. This chain MUST include the certificate or precertificate itself, the zero or more intermediate CA certificates provided by the submitter, and the root certificate used to verify the chain (even if it was omitted from the submission). The log MUST present this chain for auditing upon request (see Section 6.7). This chain is required to prevent a CA from avoiding blame by logging a partial or empty chain.
Each certificate entry in a log MUST include a X509ChainEntry structure, and each precertificate entry MUST include a PrecertChainEntryV2 structure:
opaque ASN.1Cert<1..2^24-1>;
struct {
ASN.1Cert leaf_certificate;
ASN.1Cert certificate_chain<0..2^24-1>;
} X509ChainEntry;
opaque CMSPrecert<1..2^24-1>;
struct {
CMSPrecert pre_certificate;
ASN.1Cert precertificate_chain<1..2^24-1>;
} PrecertChainEntryV2;
leaf_certificate is a submitted certificate that has been accepted by the log.
certificate_chain is a vector of 0 or more additional certificates required to verify leaf_certificate. The first certificate MUST certify leaf_certificate. Each following certificate MUST directly certify the one preceding it. The final certificate MUST be a root certificate accepted by the log. If leaf_certificate is a root certificate, then this vector is empty.
pre_certificate is a submitted precertificate that has been accepted by the log.
precertificate_chain is a vector of 1 or more additional certificates required to verify pre_certificate. The first certificate MUST certify pre_certificate. Each following certificate MUST directly certify the one preceding it. The final certificate MUST be a root certificate accepted by the log.
Each log's operator allocates an OID for the purpose of uniquely identifying that log. This OID is specified in the log's metadata. Various data structures include the DER encoding of this OID, excluding the ASN.1 tag and length bytes, in an opaque vector:
opaque LogID<2..127>;
Note that the ASN.1 length and the opaque vector length are identical in size (1 byte) and value, so the DER encoding of the OID can be reproduced simply by prepending an OBJECT IDENTIFIER tag (0x06) to the opaque vector length and contents.
Various data structures produced by logs are encapsulated in the TransItem structure to ensure that the type and version of each one is identified in a common fashion:
enum {
v1(0), v2(1), (255)
} Version;
enum {
x509_entry(0), precert_entry(1), x509_sct(2), precert_sct(3),
tree_head(4), signed_tree_head(5), consistency_proof(6),
inclusion_proof(7), (65535)
} TransType;
enum {
reserved(65535)
} ItemExtensionType;
struct {
ItemExtensionType item_extension_type;
opaque item_extension_data<0..2^16-1>;
} ItemExtension;
struct {
TransType type;
select (type) {
case x509_entry: TimestampedCertificateEntryDataV2;
case precert_entry: TimestampedCertificateEntryDataV2;
case x509_sct: SignedCertificateTimestampDataV2;
case precert_sct: SignedCertificateTimestampDataV2;
case tree_head: TreeHeadDataV2;
case signed_tree_head: SignedTreeHeadDataV2;
case consistency_proof: ConsistencyProofDataV2;
case inclusion_proof: InclusionProofDataV2;
} data;
ItemExtension item_extensions<0..2^16-1>;
} TransItemV2;
struct {
Version version;
select (version) {
case v1: TransItemV1;
case v2: TransItemV2;
}
} TransItem;
version is the earliest version of this protocol to which the encapsulated data structure conforms. This document is v2. Note that v1 [RFC6962] did not define TransItem, but this document specifies a mechanism (see Appendix A) for v2 implementations to encapsulate existing v1 objects in the TransItem structure. Note also that, since each TransItem object is individually versioned, future revisions to this protocol could conceivably update some encapsulated data structures without having to update all of them.
type is the type of the encapsulated data structure. (Note that TransType combines the v1 type enumerations LogEntryType, SignatureType and MerkleLeafType). Future revisions of this protocol may add new TransType values.
data is the encapsulated data structure. The various structures named with the DataV2 suffix are defined in later sections of this document.
item_extension_type identifies a single extension from the IANA registry in Section 11.3.
The interpretation of the item_extension_data field is determined solely by the value of the item_extension_type field. Each document that registers a new item_extension_type must describe how to interpret the corresponding item_extension_data.
item_extensions is a vector of 0 or more item extensions. This vector MUST NOT include more than one extension with the same item_extension_type. The extensions in the vector MUST be ordered by the value of the item_extension_type field, smallest value first.
The leaves of a log's Merkle Tree correspond to the log's entries (see Section 5.2). Each leaf is the leaf hash [mht] of a TransItem structure of type x509_entry or precert_entry, which in this version (v2) encapsulates a TimestampedCertificateEntryDataV2 structure. Note that leaf hashes are calculated as HASH(0x00 || TransItem), where the hashing algorithm is specified in the log's metadata.
opaque TBSCertificate<1..2^24-1>;
struct {
uint64 timestamp;
opaque issuer_key_hash[HASH_SIZE];
TBSCertificate tbs_certificate;
SctExtension sct_extensions<0..2^16-1>;
} TimestampedCertificateEntryDataV2;
timestamp is the NTP Time [RFC5905] at which the certificate or precertificate was accepted by the log, measured in milliseconds since the epoch (January 1, 1970, 00:00), ignoring leap seconds.
issuer_key_hash is the HASH of the public key of the CA that issued the certificate or precertificate, calculated over the DER encoding of the key represented as SubjectPublicKeyInfo [RFC5280]. This is needed to bind the CA to the certificate or precertificate, making it impossible for the corresponding SCT to be valid for any other certificate or precertificate whose TBSCertificate matches tbs_certificate.
tbs_certificate is the DER encoded TBSCertificate from either the leaf_certificate (in the case of an X509ChainEntry) or the pre_certificate (in the case of a PrecertChainEntryV2). (Note that a precertificate's TBSCertificate can be reconstructed from the issued certificate's TBSCertificate by redacting the domain name labels indicated by the redacted labels extension, and deleting the SCT list extension and redacted labels extension).
sct_extensions matches the SCT extensions of the corresponding SCT.
An SCT is a TransItem structure of type x509_sct or precert_sct, which in this version (v2) encapsulates a SignedCertificateTimestampDataV2 structure:
enum {
reserved(65535)
} SctExtensionType;
struct {
SctExtensionType sct_extension_type;
opaque sct_extension_data<0..2^16-1>;
} SctExtension;
struct {
LogID log_id;
uint64 timestamp;
SctExtension sct_extensions<0..2^16-1>;
digitally-signed struct {
TransItem timestamped_entry;
} signature;
} SignedCertificateTimestampDataV2;
log_id is this log's unique ID, encoded in an opaque vector as described in Section 5.3.
timestamp is equal to the timestamp from the TimestampedCertificateEntryDataV2 structure encapsulated in the timestamped_entry.
sct_extension_type identifies a single extension from the IANA registry in Section 11.4. At the time of writing, no extensions are specified.
The interpretation of the sct_extension_data field is determined solely by the value of the sct_extension_type field. Each document that registers a new sct_extension_type must describe how to interpret the corresponding sct_extension_data.
sct_extensions is a vector of 0 or more SCT extensions. This vector MUST NOT include more than one extension with the same sct_extension_type. The extensions in the vector MUST be ordered by the value of the sct_extension_type field, smallest value first. If an implementation sees an extension that it does not understand, it SHOULD ignore that extension. Furthermore, an implementation MAY choose to ignore any extension(s) that it does understand.
The encoding of the digitally-signed element is defined in [RFC5246].
timestamped_entry is a TransItem structure that MUST be of type x509_entry or precert_entry (see Section 5.5) and MUST have an empty item_extensions vector.
The log stores information about its Merkle Tree in a TransItem structure of type tree_head, which in this version (v2) encapsulates a TreeHeadDataV2 structure:
opaque NodeHash[HASH_SIZE];
struct {
uint64 timestamp;
uint64 tree_size;
NodeHash root_hash;
SthExtension sth_extensions<0..2^16-1>;
} TreeHeadDataV2;
timestamp is the current NTP Time [RFC5905], measured in milliseconds since the epoch (January 1, 1970, 00:00), ignoring leap seconds.
tree_size is the number of entries currently in the log's Merkle Tree.
root_hash is the root of the Merkle Hash Tree.
sth_extensions matches the STH extensions of the corresponding STH.
Periodically each log SHOULD sign its current tree head information (see Section 5.7) to produce an STH. When a client requests a log's latest STH (see Section 6.3), the log MUST return an STH that is no older than the log's MMD. However, STHs could be used to mark individual clients (by producing a new one for each query), so logs MUST NOT produce them more frequently than is declared in their metadata. In general, there is no need to produce a new STH unless there are new entries in the log; however, in the unlikely event that it receives no new submissions during an MMD period, the log SHALL sign the same Merkle Tree Hash with a fresh timestamp.
An STH is a TransItem structure of type signed_tree_head, which in this version (v2) encapsulates a SignedTreeHeadDataV2 structure:
enum {
reserved(65535)
} SthExtensionType;
struct {
SthExtensionType sth_extension_type;
opaque sth_extension_data<0..2^16-1>;
} SthExtension;
struct {
LogID log_id;
uint64 timestamp;
uint64 tree_size;
NodeHash root_hash;
SthExtension sth_extensions<0..2^16-1>;
digitally-signed struct {
TransItem merkle_tree_head;
} signature;
} SignedTreeHeadDataV2;
log_id is this log's unique ID, encoded in an opaque vector as described in Section 5.3.
timestamp is equal to the timestamp from the TreeHeadDataV2 structure encapsulated in merkle_tree_head. This timestamp MUST be at least as recent as the most recent SCT timestamp in the tree. Each subsequent timestamp MUST be more recent than the timestamp of the previous update.
tree_size is equal to the tree size from the TreeHeadDataV2 structure encapsulated in merkle_tree_head.
root_hash is equal to the root hash from the TreeHeadDataV2 structure encapsulated in merkle_tree_head.
sth_extension_type identifies a single extension from the IANA registry in Section 11.5. At the time of writing, no extensions are specified.
The interpretation of the sth_extension_data field is determined solely by the value of the sth_extension_type field. Each document that registers a new sth_extension_type must describe how to interpret the corresponding sth_extension_data.
sth_extensions is a vector of 0 or more STH extensions. This vector MUST NOT include more than one extension with the same sth_extension_type. The extensions in the vector MUST be ordered by the value of the sth_extension_type field, smallest value first. If an implementation sees an extension that it does not understand, it SHOULD ignore that extension. Furthermore, an implementation MAY choose to ignore any extension(s) that it does understand.
merkle_tree_head is a TransItem structure that MUST be of type tree_head (see Section 5.7) and MUST have an empty item_extensions vector.
To prepare a Merkle Consistency Proof for distribution to clients, the log produces a TransItem structure of type consistency_proof, which in this version (v2) encapsulates a ConsistencyProofDataV2 structure:
struct {
LogID log_id;
uint64 tree_size_1;
uint64 tree_size_2;
NodeHash consistency_path<1..2^8-1>;
} ConsistencyProofDataV2;
log_id is this log's unique ID, encoded in an opaque vector as described in Section 5.3.
tree_size_1 is the size of the older tree.
tree_size_2 is the size of the newer tree.
consistency_path is a vector of Merkle Tree nodes proving the consistency of two STHs.
To prepare a Merkle Inclusion Proof for distribution to clients, the log produces a TransItem structure of type inclusion_proof, which in this version (v2) encapsulates an InclusionProofDataV2 structure:
struct {
LogID log_id;
uint64 tree_size;
uint64 leaf_index;
NodeHash inclusion_path<1..2^8-1>;
} InclusionProofDataV2;
log_id is this log's unique ID, encoded in an opaque vector as described in Section 5.3.
tree_size is the size of the tree on which this inclusion proof is based.
leaf_index is the 0-based index of the log entry corresponding to this inclusion proof.
inclusion_path is a vector of Merkle Tree nodes proving the inclusion of the chosen certificate or precertificate.
Messages are sent as HTTPS GET or POST requests. Parameters for POSTs and all responses are encoded as JavaScript Object Notation (JSON) objects [RFC4627]. Parameters for GETs are encoded as order-independent key/value URL parameters, using the "application/x-www-form-urlencoded" format described in the "HTML 4.01 Specification" [HTML401]. Binary data is base64 encoded [RFC4648] as specified in the individual messages.
Note that JSON objects and URL parameters may contain fields not specified here. These extra fields should be ignored.
The <log server> prefix, which is part of the log's metadata, MAY include a path as well as a server name and a port.
In practice, log servers may include multiple front-end machines. Since it is impractical to keep these machines in perfect sync, errors may occur that are caused by skew between the machines. Where such errors are possible, the front-end will return additional information (as specified below) making it possible for clients to make progress, if progress is possible. Front-ends MUST only serve data that is free of gaps (that is, for example, no front-end will respond with an STH unless it is also able to prove consistency from all log entries logged within that STH).
For example, when a consistency proof between two STHs is requested, the front-end reached may not yet be aware of one or both STHs. In the case where it is unaware of both, it will return the latest STH it is aware of. Where it is aware of the first but not the second, it will return the latest STH it is aware of and a consistency proof from the first STH to the returned STH. The case where it knows the second but not the first should not arise (see the "no gaps" requirement above).
If the log is unable to process a client's request, it MUST return an HTTP response code of 4xx/5xx (see [RFC2616]), and, in place of the responses outlined in the subsections below, the body SHOULD be a JSON structure containing at least the following field:
e.g. In response to a request of /ct/v2/get-entries?start=100&end=99, the log would return a 400 Bad Request response code with a body similar to the following:
{
"error_message": "'start' cannot be greater than 'end'",
"error_code": "not compliant",
}
Clients SHOULD treat 500 Internal Server Error and 503 Service Unavailable responses as transient failures and MAY retry the same request without modification at a later date. Note that as per [RFC2616], in the case of a 503 response the log MAY include a Retry-After: header in order to request a minimum time for the client to wait before retrying the request.
POST https://<log server>/ct/v2/add-chain
If the version of sct is not v2, then a v2 client may be unable to verify the signature. It MUST NOT construe this as an error. This is to avoid forcing an upgrade of compliant v2 clients that do not use the returned SCTs.
If a log detects bad encoding in a chain that otherwise verifies correctly then the log MAY still log the certificate but SHOULD NOT return an SCT. It should instead return the "bad certificate" error. Logging the certificate is useful, because monitors [monitor] can then detect these encoding errors, which may be accepted by some TLS clients.
Note that not all certificate handling software is capable of detecting all encoding errors (e.g. some software will accept BER instead of DER encodings in certificates, or incorrect character encodings, even though these are technically incorrect) .
POST https://<log server>/ct/v2/add-pre-chain
Errors are the same as in Section 6.1.
GET https://<log server>/ct/v2/get-sth
No inputs.
GET https://<log server>/ct/v2/get-sth-consistency
See Section 9.4.2 for an outline of how to use the consistency output.
GET https://<log server>/ct/v2/get-proof-by-hash
See Section 9.4.1 for an outline of how to use the inclusion output.
GET https://<log server>/ct/v2/get-all-by-hash
Errors are the same as in Section 6.5.
See Section 9.4.1 for an outline of how to use the inclusion output, and see Section 9.4.2 for an outline of how to use the consistency output.
GET https://<log server>/ct/v2/get-entries
Note that this message is not signed -- the entries data can be verified by constructing the Merkle Tree Hash corresponding to a retrieved STH. All leaves MUST be v1 or v2. However, a compliant v1 client MUST NOT construe an unrecognized LogEntryType value as an error. This means it may be unable to parse some entries, but note that each client can inspect the entries it does recognize as well as verify the integrity of the data by treating unrecognized leaves as opaque input to the tree.
The start and end parameters SHOULD be within the range 0 <= x < tree_size as returned by get-sth in Section 6.3.
The start parameter MUST be less than or equal to the end parameter.
Log servers MUST honor requests where 0 <= start < tree_size and end >= tree_size by returning a partial response covering only the valid entries in the specified range. end >= tree_size could be caused by skew. Note that the following restriction may also apply:
Logs MAY restrict the number of entries that can be retrieved per get-entries request. If a client requests more than the permitted number of entries, the log SHALL return the maximum number of entries permissible. These entries SHALL be sequential beginning with the entry specified by start.
Because of skew, it is possible the log server will not have any entries between start and end. In this case it MUST return an empty entries array.
In any case, the log server MUST return the latest STH it knows about.
See Section 9.4.3 for an outline of how to use a complete list of leaf_input entries to verify the root_hash.
GET https://<log server>/ct/v2/get-roots
No inputs.
TLS servers MUST use at least one of the three mechanisms listed below to present one or more SCTs or inclusion proofs from one or more logs to each TLS client during TLS handshakes, where each SCT or inclusion proof corresponds to the server certificate or to a name-constrained intermediate the server certificate chains to. Three mechanisms are provided because they have different tradeoffs.
TLS servers SHOULD send SCTs or inclusion proofs from multiple logs in case one or more logs are not acceptable to the TLS client (for example, if a log has been struck off for misbehavior, has had a key compromise, or is not known to the TLS client).
Multiple SCTs, inclusion proofs, and indeed TransItem structures of any type, are combined into a list as follows:
opaque SerializedTransItem<1..2^16-1>;
struct {
SerializedTransItem trans_item_list<1..2^16-1>;
} TransItemList;
Here, SerializedTransItem is an opaque byte string that contains the serialized TransItem structure. This encoding ensures that TLS clients can decode each TransItem individually (so, for example, if there is a version upgrade, out-of-date clients can still parse old TransItem structures while skipping over new TransItem structures whose versions they don't understand).
TODO: We need to define at least one ItemExtensionType for associating SCT and inclusion proof TransItems with the relevant certificate.
If a TLS client includes the transparency_info extension type in the ClientHello, the TLS server MAY include the transparency_info extension in the ServerHello with extension_data set to a TransItemList. The TLS server is not expected to process or include this extension when a TLS session is resumed, since session resumption uses the original session information.
One or more TransItem structures can be embedded in the Transparency Information X.509v3 extension, which has OID <TBD> and SHOULD be non-critical. This extension can be included in OCSP responses and certificates. Since RFC5280 requires the extnValue field (an OCTET STRING) of each X.509v3 extension to include the DER encoding of an ASN.1 value, we cannot embed a TransItemList directly. Instead, we have to wrap it inside an additional OCTET STRING, which we then put into the extnValue field:
TransparencyInformationSyntax ::= OCTET STRING
TransparencyInformationSyntax contains a TransItemList.
A certification authority may include a Transparency Information X.509v3 extension in the singleExtensions of a SingleResponse in an OCSP response. The included SCTs or inclusion proofs MUST be for the certificate identified by the certID of that SingleResponse, or for a precertificate that corresponds to that certificate, or for a name-constrained intermediate to which that certificate chains.
A certification authority may include a Transparency Information X.509v3 extension in a certificate. Any included SCTs or inclusion proofs MUST be either for a precertificate that corresponds to this certificate, or for a name-constrained intermediate to which this certificate chains.
There are various different functions clients of logs might perform. We describe here some typical clients and how they should function. Any inconsistency may be used as evidence that a log has not behaved correctly, and the signatures on the data structures prevent the log from denying that misbehavior.
All clients need various metadata in order to communicate with logs and verify their responses. This metadata is described below, but note that this document does not describe how the metadata is obtained, which is implementation dependent (see, for example, [Chromium.Policy]).
Clients should somehow exchange STHs they see, or make them available for scrutiny, in order to ensure that they all have a consistent view. The exact mechanisms will be in separate documents, but it is expected there will be a variety.
In order to communicate with and verify a log, clients need metadata about the log.
[JSON.Metadata] is an example of a metadata format which includes the above elements.
TLS clients receive SCTs alongside or in certificates, either for the server certificate itself or for a name-constrained intermediate the server certificate chains to. TLS clients MUST implement all of the three mechanisms by which TLS servers may present SCTs (see Section 7). TLS clients that support the transparency_info TLS extension SHOULD include it in ClientHello messages, with extension_data set to <TBD>.
TODO: What should the TLS client communicate in the extension_data? Version(s) of CT that it supports? Certain types of TransItem that it can handle? Whether or not it wants to gossip?
In addition to normal validation of the certificate and its chain, TLS clients SHOULD validate each supplied SCT by computing the signature input from the SCT data as well as the certificate and verifying the signature, using the corresponding log's public key. TLS clients MUST reject SCTs whose timestamp is in the future.
TLS clients SHOULD also validate each supplied inclusion proof (see Section 9.4.1), in order to audit the log. If no inclusion proof was supplied by the TLS server, the TLS client MAY request one directly from the corresponding log using get-proof-by-hash (Section 6.5) or get-all-by-hash (Section 6.6), and then validate it.
To be considered compliant, a certificate MUST be accompanied by at least one valid SCT or at least one valid inclusion proof. A certificate not accompanied by any valid SCTs or any valid inclusion proofs MUST NOT be considered compliant by TLS clients. However, specifying the TLS clients' behavior once compliance or non-compliance has been determined (for example, whether a certificate should be rejected due to non-compliance) is outside the scope of this document.
If the TLS client holds an STH that predates the SCT, it MAY, in the process of auditing, request a new STH from the log (Section 6.3), then verify it by requesting a consistency proof (Section 6.4). Note that if the TLS client uses get-all-by-hash, then it will already have the new STH.
Monitors watch logs and check that they behave correctly. Monitors may additionally watch for certificates of interest. For example, a monitor may be configured to report on all certificates that apply to a specific domain name when fetching new entries for consistency validation.
A monitor needs to, at least, inspect every new entry in each log it watches. It may also want to keep copies of entire logs. In order to do this, it should follow these steps for each log:
Or, if it is not keeping all log entries:
Auditing is taking partial information about a log as input and verifying that this information is consistent with other partial information held. All clients described above may perform auditing as an additional function. The action taken by the client if audit fails is not specified, but note that in general if audit fails, the client is in possession of signed proof of the log's misbehavior.
A monitor [monitor] can audit by verifying the consistency of STHs it receives, ensure that each entry can be fetched and that the STH is indeed the result of making a tree from all fetched entries.
A TLS client [tls_clients] can audit by verifying an SCT against any STH dated after the SCT timestamp + the Maximum Merge Delay by requesting a Merkle inclusion proof (Section 6.5). It can also verify that the SCT corresponds to the certificate it arrived with (i.e. the log entry is that certificate, is a precertificate for that certificate or is an appropriate name-constrained intermediate [see Section 4.3]).
The following algorithm outlines may be useful for clients that wish to perform various audit operations.
When a client has received a TransItem of type inclusion_proof and wishes to verify inclusion of an input hash for an STH with a given tree_size and root_hash, the following algorithm may be used to prove the hash was included in the root_hash:
Otherwise:
When a client has an STH first_hash for tree size first, an STH second_hash for tree size second where 0 < first < second, and has received a TransItem of type consistency_proof that they wish to use to verify both hashes, the following algorithm may be used:
Otherwise:
When a client has a complete list of leaf input entries from 0 up to tree_size - 1 and wishes to verify this list against an STH root_hash returned by the log for the same tree_size, the following algorithm may be used:
It is not possible for a log to change any of its algorithms part way through its lifetime. If it should become necessary to deprecate an algorithm used by a live log, then the log should be frozen as specified in Section 9.1 and a new log should be started. If necessary, the new log can contain existing entries from the frozen log, which monitors can verify are an exact match.
IANA is asked to allocate an RFC 5246 ExtensionType value for the transparency_info TLS extension. IANA should update this extension type to point at this document.
IANA is asked to establish a registry of hash values, initially consisting of:
| Index | Hash |
|---|---|
| 0 | SHA-256 [FIPS.180-4] |
IANA is asked to establish a registry of TransItem extensions, initially consisting of:
| Type | Extension |
|---|---|
| 65535 | reserved |
TBD: policy for adding to the registry
IANA is asked to establish a registry of SCT extensions, initially consisting of:
| Type | Extension |
|---|---|
| 65535 | reserved |
TBD: policy for adding to the registry
IANA is asked to establish a registry of STH extensions, initially consisting of:
| Type | Extension |
|---|---|
| 65535 | reserved |
TBD: policy for adding to the registry
With CAs, logs, and servers performing the actions described here, TLS clients can use logs and signed timestamps to reduce the likelihood that they will accept misissued certificates. If a server presents a valid signed timestamp for a certificate, then the client knows that a log has committed to publishing the certificate. From this, the client knows that the subject of the certificate has had some time to notice the misissue and take some action, such as asking a CA to revoke a misissued certificate, or that the log has misbehaved, which will be discovered when the SCT is audited. A signed timestamp is not a guarantee that the certificate is not misissued, since the subject of the certificate might not have checked the logs or the CA might have refused to revoke the certificate.
In addition, if TLS clients will not accept unlogged certificates, then site owners will have a greater incentive to submit certificates to logs, possibly with the assistance of their CA, increasing the overall transparency of the system.
Misissued certificates that have not been publicly logged, and thus do not have a valid SCT, are not considered compliant (so TLS clients may decide, for example, to reject them). Misissued certificates that do have an SCT from a log will appear in that public log within the Maximum Merge Delay, assuming the log is operating correctly. Thus, the maximum period of time during which a misissued certificate can be used without being available for audit is the MMD.
The logs do not themselves detect misissued certificates; they rely instead on interested parties, such as domain owners, to monitor them and take corrective action when a misissue is detected.
CAs SHOULD NOT redact domain name labels in precertificates such that the entirety of the domain space below the unredacted part of the domain name is not owned or controlled by a single entity (e.g. ?.com and ?.co.uk would both be problematic). Logs MUST NOT reject any precertificate that is overly redacted but which is otherwise considered compliant. It is expected that monitors will treat overly redacted precertificates as potentially misissued. TLS clients MAY reject a certificate whose corresponding precertificate would be overly redacted, perhaps using the same mechanism for determining whether a wildcard in a domain name of a certificate is too broad.
A log can misbehave in two ways: (1) by failing to incorporate a certificate with an SCT in the Merkle Tree within the MMD and (2) by violating its append-only property by presenting two different, conflicting views of the Merkle Tree at different times and/or to different parties. Both forms of violation will be promptly and publicly detectable.
Violation of the MMD contract is detected by log clients requesting a Merkle audit proof for each observed SCT. These checks can be asynchronous and need only be done once per each certificate. In order to protect the clients' privacy, these checks need not reveal the exact certificate to the log. Clients can instead request the proof from a trusted auditor (since anyone can compute the audit proofs from the log) or request Merkle proofs for a batch of certificates around the SCT timestamp.
Violation of the append-only property can be detected by clients comparing their instances of the Signed Tree Heads. As soon as two conflicting Signed Tree Heads for the same log are detected, this is cryptographic proof of that log's misbehavior. There are various ways this could be done, for example via gossip (see http://trac.tools.ietf.org/id/draft-linus-trans-gossip-00.txt) or peer-to-peer communications or by sending STHs to monitors (who could then directly check against their own copy of the relevant log).
TLS servers may wish to offer multiple SCTs, each from a different log.
The Merkle Tree design serves the purpose of keeping communication overhead low.
Auditing logs for integrity does not require third parties to maintain a copy of each entire log. The Signed Tree Heads can be updated as new entries become available, without recomputing entire trees. Third-party auditors need only fetch the Merkle consistency proofs against a log's existing STH to efficiently verify the append-only property of updates to their Merkle Trees, without auditing the entire tree.
The authors would like to thank Erwann Abelea, Robin Alden, Al Cutter, Francis Dupont, Adam Eijdenberg, Stephen Farrell, Daniel Kahn Gillmor, Brad Hill, Jeff Hodges, Paul Hoffman, Jeffrey Hutzelman, Stephen Kent, SM, Alexey Melnikov, Linus Nordberg, Chris Palmer, Trevor Perrin, Pierre Phaneuf, Melinda Shore, Ryan Sleevi, Carl Wallace and Paul Wouters for their valuable contributions.
| [Chromium.Log.Policy] | The Chromium Projects, "Chromium Certificate Transparency Log Policy", 2014. |
| [Chromium.Policy] | The Chromium Projects, "Chromium Certificate Transparency", 2014. |
| [CrosbyWallach] | Crosby, S. and D. Wallach, "Efficient Data Structures for Tamper-Evident Logging", Proceedings of the 18th USENIX Security Symposium, Montreal, August 2009. |
| [EVSSLGuidelines] | CA/Browser Forum, "Guidelines For The Issuance And Management Of Extended Validation Certificates", 2007. |
| [JSON.Metadata] | The Chromium Projects, "Chromium Log Metadata JSON Schema", 2014. |
| [RFC6962] | Laurie, B., Langley, A. and E. Kasper, "Certificate Transparency", RFC 6962, DOI 10.17487/RFC6962, June 2013. |
TODO: Finish writing this section. Or should it be in a separate document?
struct {
TransType type;
select (type) {
case x509_sct: SignedCertificateTimestampV1;
case precert_sct: SignedCertificateTimestampV1;
case signed_tree_head: SignedTreeHeadDataV1;
case consistency_proof: ConsistencyProofDataV1;
case inclusion_proof: InclusionProofDataV1;
} data;
} TransItemV1;
opaque SHA256Hash[32];
struct {
Version version = v1;
SHA256Hash log_id;
uint64 timestamp;
SctExtensions extensions;
digitally-signed struct {
Version version = v1;
uint8 signature_type = 0; /* "certificate_timestamp" */
uint64 timestamp;
TransType type; /* x509_entry(0) or precert_entry(1) */
select (type) {
case x509_entry: ASN.1Cert;
case precert_entry: PreCert;
} signed_entry;
SctExtensions extensions;
} signature;
} SignedCertificateTimestampV1;
struct {
SHA256Hash log_id;
uint64 timestamp;
uint64 tree_size;
SHA256Hash sha256_root_hash;
digitally-signed struct {
Version version = v1;
uint8 signature_type = 1; /* "tree_hash" */
uint64 timestamp;
uint64 tree_size;
SHA256Hash sha256_root_hash;
} signature;
} SignedTreeHeadDataV1;
struct {
SHA256Hash log_id;
uint64 tree_size_1;
uint64 tree_size_2;
SHA256Hash consistency_path<1..2^8-1>;
} ConsistencyProofDataV1;
struct {
SHA256Hash log_id;
uint64 tree_size;
uint64 leaf_index;
SHA256Hash inclusion_path<1..2^8-1>;
} InclusionProofDataV1;