Commercial National Security Algorithm (CNSA) Suite Profile for Secure/Multipurpose Internet Mail Extensions (S/MIME)

Abstract

This document defines a base profile of S/MIME for use with the US Commercial National Security Algorithm (CNSA) 2.0 Suite, a cybersecurity advisory published by the United States Government which outlines quantum-resistant cryptographic algorithm policy for US national security applications.¶

This profile applies to the capabilities, configuration, and operation of all components of US National Security Systems that employ S/MIME. It is also appropriate for all other US Government systems that process high-value information.¶

This memo is not an IETF standard, and has not been shown to have IETF community consensus. This profile is made publicly available for use by developers and operators of these and any other system deployments.¶

1. Introduction

This document specifies a profile of S/MIME [RFC8551] to comply with the National Security Agency's (NSA) Commercial National Security Algorithm (CNSA) 2.0 Suite [cnsafaq]. This profile applies to the capabilities, configuration, and operation of all components of US National Security Systems (NSS) [SP80059] that employ S/MIME. This profile is also appropriate for all other US Government systems that process high-value information, and is made publicly available for use by developers and operators of these and any other system deployments.¶

This document does not define any cryptographic algorithm, and does not specify how to use any cryptographic algorithm not currently supported by S/MIME; instead, it profiles CNSA 2.0-compliant conventions for S/MIME, and uses algorithms already specified for use in S/MIME that are also allowed by the CNSA Suite 2.0. This document applies to all CNSA Suite solutions that make use of S/MIME.¶

The reader is assumed to have familiarity with [RFC8551].¶

This memo is not an IETF standard, and has not been shown to have IETF community consensus.¶

[ED NOTE: This document uses some details from [I-D.ietf-lamps-cms-kyber] to specify use of ML-KEM in CMS, and [I-D.ietf-lamps-cms-ml-dsa] to specify use of ML-DSA in CMS. We note that these drafts are not yet RFCs, and we may need to adjust this text accordingly based on the progress of these documents.]¶

2. Terminology

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 [RFC2119] [RFC8174] when, and only when, they appear in all capitals, as shown here. Normative language does not apply beyond the scope of this profile.¶

This is a profile of S/MIME [RFC8551]. Therefore, the requirements language in this memo may be different than that found in the underlying standards.¶

All references to "CNSA" in this document refer to CNSA 2.0 [cnsafaq], unless otherwise stated.¶

3. The Commercial National Security Algorithm Suite

The CNSA Suite is the set of approved commercial algorithms that can be used by vendors and IT users to meet cybersecurity and interoperability requirements for NSS. The initial suite of CNSA Suite algorithms, Suite B, established a baseline for use of commercial algorithms to protect classified information. The following suite, CNSA 1.0, served as a bridge between the original set and a fully quantum-resistant cryptographic capability. The current suite, CNSA 2.0, seeks to provide fully quantum-resistant protection [cnsafaq].¶

This profile uses CNSA 2.0 compliant algorithms: ML-DSA-87 [FIPS204] for signing, and ML-KEM-1024 [FIPS203] for key establishment. ML-DSA-87 and ML-KEM-1024 together with SHA384, SHA512, AES-256, LMS, and XMSS comprise the CNSA Suite 2.0.¶

The NSA is authoring a set of RFCs, including this one, to provide updated guidance for using the CNSA 2.0 Suite in certain IETF protocols. These RFCs can be used in conjunction with other RFCs and cryptographic guidance (e.g., NIST Special Publications) to properly protect Internet traffic and data-at-rest for US Government NSS.¶

4. Requirements and Assumptions

CMS values are generated using ASN.1 [X208], the Basic Encoding Rules (BER) [X209], and the Distinguished Encoding Rules (DER) [X509].¶

For key agreement, CNSA compliant implementations MUST use ML-KEM-1024. Details provided in Section 7.¶

For digital signature, CNSA compliant implementations MUST use ML-DSA-87.¶

For CNSA Suite applications, public key certificates used to verify S/MIME signatures MUST be compliant with the CNSA Suite Certificate and Certificate Revocation List (CRL) profile specified in [I-D.jenkins-cnsa2-pkix-profile].¶

Within the CMS signed-data content type, signature algorithm identifiers are located in the signatureAlgorithm field of SignerInfo structures contained within the SignedData. In addition, signature algorithm identifiers are located in the SignerInfo signatureAlgorithm field of countersignature attributes. Specific requirements for digital signatures are given in Section 6; compliant implementations MUST consider signatures not meeting these requirements as invalid.¶

This document assumes that the required trust anchors have been securely provisioned to the client.¶

All implementations MUST use SHA-384 or SHA-512 for hashing and AES-GCM for encryption, the requirements for which are given in the following sections.¶

5. Message Digest

CNSA-compliant implementations MUST use either SHA-384 or SHA-512 for message digest.¶

[RFC5754] specifies the conventions for using SHA-384 or SHA-512 with the Cryptographic Message Syntax (CMS). CNSA-compliant S/MIME implementations MUST follow the conventions in [RFC5754].¶

Within the CMS signed-data content type, message digest algorithm identifiers are located in the SignedData digestAlgorithms field and the SignerInfo digestAlgorithm field.¶

The SHA-384 message digest algorithm is defined in FIPS 180 [SHS], and the algorithm identifier for SHA-384 is defined in [RFC5754] as follows:¶

id-sha384 OBJECT IDENTIFIER ::= {¶

joint-iso-itu-t(2) country(16) us(840) organization(1) gov(101)¶

csor(3) nistalgorithm(4) hashalgs(2) 2 }¶

The SHA-512 message digest algorithm is defined in FIPS 180 [SHS], and the algorithm identifier for SHA-512 is defined in [RFC5754] as follows:¶

id-sha512 OBJECT IDENTIFIER ::= {¶

joint-iso-itu-t(2) country(16) us(840) organization(1) gov(101)¶

csor(3) nistalgorithm(4) hashalgs(2) 3 }¶

[RFC5754] defines the AlgorithmIdentifier parameters field as OPTIONAL. Implementations MUST accept SHA-384 and SHA-512 AlgorithmIdentifiers with absent parameters, or with NULL parameters. As specified in [RFC5754], implementations MUST generate SHA-384 or SHA-512 AlgorithmIdentifiers with absent parameters.¶

6. Digital Signature

CNSA-compliant implementations MUST use ML-DSA-87 as the digital signature algorithm in S/MIME.¶

[ED NOTE: The algorithm identifier for ML-DSA-87 is defined in Section 2 of [I-D.ietf-lamps-cms-ml-dsa], which is ongoing work detailing the conventions for using ML-DSA with the Cryptographic Message Syntax (CMS); This document will be updated accordingly as that work progresses. Note that [I-D.ietf-lamps-cms-ml-dsa] only specifies use of the pure mode (not pre-hash) of ML-DSA in CMS, which is also in accordance with CNSA. The "message digest" supplied to the signature algorithm is the entire message.¶

[EDNOTE: This guidance is subject to change as conventions for External Mu, as defined in [I-D.ietf-lamps-dilithium-certificates], are further defined.]¶

7. Key Establishment

8. Content Encryption

CNSA-compliant S/MIME implementations using the authenticated-enveloped-data content type [RFC5083] MUST use AES [AES] in Galois Counter Mode (GCM) [SP80038D] as the content authenticated encryption algorithm and MUST follow the conventions for using AES-GCM with the CMS defined in [RFC5084].¶

Within the CMS authenticated-enveloped-data content type, content-authenticated encryption algorithm identifiers are located in the AuthEnvelopedData EncryptedContentInfo contentEncryptionAlgorithm field. The content-authenticated encryption algorithm is used to encipher the content located in the AuthEnvelopedData EncryptedContentInfo encryptedContent field.¶

The AES-GCM content-authenticated encryption algorithm is described in [GCM] and [SP80038D]. The algorithm identifier for AES-256 in GCM mode is:¶

id-aes256-GCM OBJECT IDENTIFIER ::= { joint-iso-itu-t(2)¶

country(16) us(840) organization(1) gov(101) csor(3)¶

nistAlgorithm(4) aes(1) 46 }¶

The AlgorithmIdentifier parameters field MUST be present, and the parameters field must contain GCMParameters:¶

GCMParameters ::= SEQUENCE { )¶

aes-nonce OCTET STRING,¶

aes-ICVlen AES-GCM-ICVlen DEFAULT 12 }¶

The authentication tag length (aes-ICVlen) SHALL be 16 (indicating a tag length of 128 bits).¶

The initialization vector (aes-nonce) MUST be generated in accordance with Section 8.2 of [SP80038D]. AES-GCM loses security catastrophically if a nonce is reused with a given key on more than one distinct set of input data. Therefore, a fresh content-authenticated encryption key MUST be generated for each message.¶

9. Security Considerations

Most of the security considerations for this document are described in [RFC8551] and [RFC9629]. Additional security considerations for the use of ML-KEM and ML-DSA in S/MIME can be found in [I-D.ietf-lamps-cms-kyber] and [I-D.ietf-lamps-cms-ml-dsa], respectively.¶

11. References