Countersigning COSE Envelopes in Transparency Services

Introduction This document defines a method for issuing and verifying countersignatures on COSE_Sign1 messages included in an authenticated data structure such as a Merkle Tree. We adopt the terminology of architecture for Claim, Envelope, Transparency Service, Ledger, Receipt, and Verifier.

[TODO] Do we need to explain or introduce them here? We may also define Tree (our shorthand for authenticated data structure), Root (a succinct commitment to the Tree, e.g., a hand) and use Issuer instead of TS.

From the Verifier's viewpoint, a Receipt is similar to a countersignature V2 on a single signed message: it is a universally-verifiable cryptographic proof of endorsement of the signed envelope by the countersigner. Compared with countersignatures on single COSE envelopes, - Receipts countersign the envelope in context, providing authentication both of the envelope and of its logical position in the authenticated data structure. - Receipts are proof of commitment to the whole contents of the data structure, even if the Verifier knows only some of its contents. - Receipts can be issued in bulk, using a single public-key signature for issuing a large number of Receipts.

Requirements Notation 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.

Common Parameters Verifiers are configured by a collection of parameters to identify a Transparency Service and verify its Receipts. These parameters MUST be fixed for the lifetime of the Transparency Service and securely communicated to all Verifiers. At minimum, these parameters include:

a Service identifier: An opaque identifier (e.g. UUID) that uniquely identifies the service and can be used to securely retrive all other Serivce parameters.
The Tree algorithm used for issuing receipts, and its additional global parameters, if any. This document creates a registry (see ) and describes an initial set of tree algorithms.
- [TODO] The architecture also has fixed TS registration policies.

Generic Receipt Structure A Receipt represents a countersignature issued by a Transparency Service. The Receipt structure is a CBOR array with two items, in order:

service_id: The service identifier as tstr.
contents: The proof as a CBOR structure determined by the tree algorithm.

Receipt = [ service_id: tstr, contents: any ] Each tree algorithm MUST define its contents type and procedures for issuing and verifying a receipt.

COSE_Sign1 Countersigning While the tree algorithms may differ in the way they aggregate multiple envelopes to compute a digest to be signed by the TS, they all share the same representation of the individual envelopes to be countersigned (intuitively, their leaves). This document uses the principals and structure definitions of COSE_Sign1 countersigning V2 (). Each envelope is authenticated using a Countersign_structure array, recalled below. Countersign_structure = [ context: "CounterSignatureV2", body_protected: empty_or_serialized_map, sign_protected: empty_or_serialized_map, external_aad: bstr, payload: bstr, other_fields: [ signature: bstr ] ] The body_protected, payload, and signature fields are copied from the COSE_Sign1 message being countersigned. The sign_protected field is provided by the TS, see below. This field is included in the Receipt contents to enable the Verifier to re-construct Countersign_structure, as specified by the tree algorithm. By convention, the TS always provides an empty external_aad: a zero-length bytestring. Procedure for reconstruction of Countersign_structure:

Let Target be the COSE_Sign1 message that corresponds to the countersignature. Different environments will have different mechanisms to achieve this. One obvious mechanism is to embed the Receipt in the unprotected header of Target. Another mechanism may be to store both artifacts separately and use a naming convention, database, or other method to link both together.
Extract body_protected, payload, and signature from Target.
Create a Countersign_structure using the extracted fields from Target, and sign_protected from the Receipt contents.

Countersigner Header Parameters The following parameters MUST be included in the protected header of the countersigner (sign_protected in ):

Issued At (label: TBD): The time at which the countersignature was issued as the number of seconds from 1970-01-01T00:00:00Z UTC, ignoring leap seconds.

CCF 2 Tree Algorithm The CCF 2 tree algorithm specifies an algorithm based on a binary Merkle tree over the sequence of all ledger entries, as implemented in the CCF version 2 framework (see ).

Additional Parameters The algorithm requires that the TS define additional parameters:

Hash Algorithm: The hash algorithm used in its Merkle Tree (see ).
Signature Algorithm: The signature algorithm used (see ).
Service Certificate: The self-signed X.509 certificate used as trust anchor to verify signatures generated by the transparency service using the Signature Algorithm.

All definitions in this section use the hash algorithm set in the TS parameters (see Section ). We write HASH to refer to this algorithm, and HASH_SIZE for the fixed length of its output in bytes.

Cryptographic Components Note: This section is adapted from , which provides additional discussion of Merkle trees.

Binary Merkle Trees The input of the Merkle Tree Hash (MTH) function is a list of n bytestrings, written D_n = {d[0], d[1], ..., d[n-1]}. The output is a single HASH_SIZE bytestring, also called the tree root hash. This function is defined as follows: The hash of an empty list is the hash of an empty string: The hash of a list with one entry (also known as a leaf hash) is: 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: where:

|| denotes concatenation
: denotes concatenation of lists
D[k1:k2] = D'_(k2-k1) denotes the list {d'[0] = d[k1], d'[1] = d[k1+1], ..., d'[k2-k1-1] = d[k2-1]} of length (k2 - k1).

Merkle Inclusion Proofs A Merkle inclusion proof for a leaf in a Merkle Tree is the shortest list of intermediate hash values required to re-compute the tree root hash form the digest of the leaf bytestring. 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.

Verifying an Inclusion Proof When a client has received an inclusion proof and wishes to verify inclusion of a leaf_hash for a given root_hash, the following algorithm may be used to prove the hash was included in the root_hash:

Generating an Inclusion Proof Given the MTH input D_n = {d[0], d[1], ..., d[n-1]} and an index i < n in this list, run the MTH algorithm and record the position and value of every intermediate hash concatenated and hashed first with the digest of the leaf, then with the resulting intermediate hash value. (Most implementations instead record all intermediate hash computations, so that they can produce all inclusion proofs for a given tree by table lookups.)

Encoding Signed Envelopes into Tree Leaves This section describes the encoding of signed envelopes and auxiliary ledger entries into the leaf bytestrings passed as input to the Merkle Tree function. Each bytestring is computed from three inputs:

internal_hash: a string of HASH_SIZE bytes;
internal_data: a string of at most 1024 bytes; and
data_hash: either the HASH of the CBOR-encoded Countersign_structure of the signed envelope, using the CBOR encoding described in , or a bytestring of size HASH_SIZE filled with zeroes for auxiliary ledger entries.

as the concatenation of three hashes: This ensures that leaf bytestrings are always distinct from the inputs of the intermediate computations in MTH, which always consist of two hashes, and also that leaf bytestrings for signed envelopes and for auxiliary ledger entries are always distinct. The internal_hash and internal_data bytestrings are internal to the CCF implementation Similarly, the auxiliary ledger entries are internal to CCF. They are opaque to receipt Verifiers, but they commit the TS to the whole ledger contents and may be used for additional, CCF-specific auditing.

Receipt Contents Structure The Receipt contents structure is a CBOR array. The items of the array in order are:

signature: the signature over the Merkle tree root as bstr.
node_certificate: a DER-encoded X.509 certificate for the public key for signature verification. This certificate MUST be a valid CCF node certificate for the service; in particular, it MUST form a valid X.509 certificate chain with the service certificate.
inclusion_proof: the intermediate hashes to recompute the signed root of the Merkle tree from the leaf digest of the envelope.
- The array MUST have at most 64 items.
- The inclusion proof structure is an array of [left, hash] pairs where left indicates the ordering of digests for the intermediate hash compution. The hash MUST be a bytestring of length HASH_SIZE.
leaf_info: auxiliary inputs to recompute the leaf digest included in the Merkle tree: the internal hash, the internal data, and the protected header of the countersigner.
- internal_hash MUST be a bytestring of length HASH_SIZE;
- internal_data MUST be a bytestring of length less than 1024.

The inclusion of an additional, short-lived certificate endorsed by the TS enables flexibility in its distributed implementation, and may support additional CCF-specific auditing. The CDDL fragment that represents the above text follows. ReceiptContents = [ signature: bstr, node_certificate: bstr, inclusion_proof: [+ ProofElement], leaf_info: LeafInfo ] ProofElement = [ left: bool hash: bstr ] LeafInfo = [ internal_hash: bstr, internal_data: bstr, sign_protected: empty_or_serialized_map ]

Receipt Verification Given the TS parameters, a signed envelope, and a Receipt for it, the following steps must be followed to verify this Receipt.

Verify that the Receipt Content structure is well-formed, as described in
Construct a Countersign_structure as described in , using sign_protected from the leaf_info field of the receipt contents.
Compute LeafBytes as the bytestring concatenation of the internal hash, the hash of internal data, and the hash of the CBOR-encoding of Countersign_structure, using the CBOR encoding described in .
Compute the leaf digest.
Compute the root hash from the leaf hash and the Merkle proof using the Merkle Tree Hash Algorithm found in the service's parameters (see ):
Verify the certificate chain established by the node certificate embedded in the receipt and the fixed service certificate in the TS parameters (see ) using the Issued At time from sign_protected to verify the validity periods of the certificates. The chain MUST enable the use of the public key in the receipt certificate for signature verification with the Signature Algorithm of the TS parameters.
Verify that signature is a valid signature value of the root hash, using the public key of the receipt certificate and the Signature Algorithm of the TS parameters.

The Verifier SHOULD apply additional checks before accepting the countersigned envelope as valid, based on its protected headers and payload.

Receipt Generation This document provides a reference algorithm for producing valid receipts, but it omits any discussion of TS registration policy and any CCF-specific implementation details. The algorithm takes as input a list of entries to be jointly countersigned, each entry consisting of internal_hash, internal_data, and an optional signed envelope. (This optional item reflects that a CCF ledger records both signed envelopes and auxiliary entries.)

For each signed envelope, compute the Countersign_structure as described in .
For each item in the list, compute LeafBytes as the bytestring concatenation of the internal hash, the hash of internal data and, if the envelope is present, the hash of the CBOR-encoding of Countersign_structure, using the CBOR encoding described in , otherwise a HASH_SIZE bytestring of zeroes.
Compute the tree root hash by applying MTH to the resulting list of leaf bytestrings, keeping the results for all intermediate HASH values.
Select a valid node_certificate and compute a signature of the root of the tree with the corresponding signing key.
For each signed envelope provided in the input,
- Collect an inclusion_proof by selecting intermediate hash values, as described above.
- Produce the receipt contents using this inclusion_proof, the fixed node_certificate and signature, and the bytestrings internal_hash and internal_data provided with the envelope.
- Produce the receipt using the Service Identifier and this receipt contents.

CBOR Encoding Restrictions In order to always regenerate the same byte string for the "to be signed" and "to be hashed" values, the core deterministic encoding rules defined in MUST be used for all their CBOR structures.

Privacy Considerations TBD

Security Considerations TBD

IANA Considerations

Additions to Existing Registries

New Entries to the COSE Header Parameters Registry IANA is requested to register the new COSE Header parameters defined below in the "COSE Header Parameters" registry.

COSE_Sign1 Countersign receipt Name: COSE_Sign1 Countersign receipt Label: TBD Value Type: Receipt / [+ Receipt] Description: A COSE_Sign1 Countersign Receipt to be embedded in the unprotected header of the countersigned COSE_Sign1 message.

Issued At Name: Issued At Label: TBD Value Type: uint Description: The time at which the signature was issued as the number of seconds from 1970-01-01T00:00:00Z UTC, ignoring leap seconds.

New SCITT-Related Registries IANA is asked to add a new registry "TBD" to the list that appears at https://www.iana.org/assignments/. The rest of this section defines the subregistries that are to be created within the new "TBD" registry.

Tree Algorithms IANA is asked to establish a registry of tree algorithm identifiers, named "Tree Algorithms", with the following registration procedures: TBD The "Tree Algorithms" registry initially consists of: Initial content of Tree Algorithms registry

Identifier	Tree Algorithm	Reference
CCF-2	CCF 2 tree algorithm	This document

The designated expert(s) should ensure that the proposed algorithm has a public specification and is suitable for use as [TBD].

Hash Algorithms IANA is asked to establish a registry of hash algorithm identifiers, named "Hash Algorithms", with the following registration procedures: TBD The "Hash Algorithms" registry initially consists of: Initial content of Hash Algorithms registry

Identifier	Hash Algorithm	Reference
SHA-256	SHA-256

The designated expert(s) should ensure that the proposed algorithm has a public specification and is suitable for use as a cryptographic hash algorithm with no known preimage or collision attacks. These attacks can damage the integrity of the ledger.

Signature Algorithms IANA is asked to establish a registry of signature algorithm identifiers, named "Signature Algorithms", with the following registration procedures: TBD The "Signature Algorithms" registry initially consists of: Initial content of Signature Algorithms registry

Identifier	Signature Algorithm	Reference
ES256	Deterministic ECDSA (NIST P-256) with HMAC-SHA256
ED25519	Ed25519 (PureEdDSA with the edwards25519 curve)

The designated expert(s) should ensure that the proposed algorithm has a public specification and is suitable for use as a cryptographic signature algorithm.