]> PQShield

Nijmegen The Netherlands thom@thomwiggers.nl

Brave Software

Lisbon Portugal cherenkov@riseup.net

Radboud University and MPI-SP

peter@cryptojedi.org

University of Waterloo

Waterloo, ON Canada dstebila@uwaterloo.ca

nicholas.sullivan+ietf@gmail.com

SECAREA TLS Working Group Internet-Draft This document gives a construction for a Key Encapsulation Mechanism (KEM)-based authentication mechanism in TLS 1.3. This proposal authenticates peers via a key exchange protocol, using their long-term (KEM) public keys. About This Document Status information for this document may be found at . Discussion of this document takes place on the tlswg Working Group mailing list (), which is archived at . Subscribe at . Source for this draft and an issue tracker can be found at .

Introduction Note: This is a work-in-progress draft. We welcome discussion, feedback and contributions through the IETF TLS working group mailing list or directly on GitHub. Any code points indicated by this draft are for experiments only, and should not be expected to be stable. This document gives a construction for KEM-based authentication in TLS 1.3 . Authentication happens via asymmetric cryptography by the usage of KEMs advertised as the long-term KEM public keys in the Certificate. TLS 1.3 is in essence a signed key exchange protocol (if using certificate-based authentication). Authentication in TLS 1.3 is achieved by signing the handshake transcript with digital signatures algorithms. KEM-based authentication provides authentication by deriving a shared secret that is encapsulated against the public key contained in the Certificate. Only the holder of the private key corresponding to the certificate's public key can derive the same shared secret and thus decrypt its peer's messages. This approach is appropriate for endpoints that have KEM public keys. Though this is currently rare, certificates can be issued with (EC)DH public keys as specified for instance in , or using a delegation mechanism, such as delegated credentials . In this proposal, we build on . This standard currently only covers Diffie-Hellman based KEMs, but the first post-quantum algorithms have already been put forward . This proposal uses ML-KEM , the first selected algorithm for key exchange in the NIST post-quantum standardization project . describes how ML-KEM keys can be encoded in X.509 certificates.

Revision history This section should be removed prior to publication of a final version of this document.

• Revision draft-celi-wiggers-tls-authkem-04
  • Some updates to ML-KEM
  • Fixed some links
• Revision draft-celi-wiggers-tls-authkem-03
  • Assigned experimental code points
  • Re-worked HPKE computation
• Revision draft-celi-wiggers-tls-authkem-02
  • Split PSK mechanism off into
  • Editing
• Revision draft-celi-wiggers-tls-authkem-01
  • Significant Editing
  • Use HPKE context
• Revision draft-celi-wiggers-tls-authkem-00
  • Initial version

Revision 2

Using key exchange instead of signatures for authentication The elliptic-curve and finite-field-based key exchange and signature algorithms that are currently widely used are very similar in sizes for public keys, ciphertexts and signatures. As an example, RSA signatures are famously "just" RSA encryption backwards. This changes in the post-quantum setting. Post-quantum key exchange and signature algorithms have significant differences in implementation, performance characteristics, and key and signature sizes. This also leads to increases in code size: For example, implementing highly efficient polynomial multiplication for post-quantum key-encapsulation mechanism ML-KEM and signature scheme ML-DSA requires significantly different approaches, even though the algorithms are related . Using the protocol proposed in this draft allows to reduce the amount of data exchanged for handshake authentication. It also allows to re-use the implementation that is used for ephemeral key exchange for authentication, as KEM operations replace signing. This decreases the code size requirements, which is especially relevant to protected implementations. Finally, KEM operations may be more efficient than signing, which might especially affect embedded platforms.

Evaluation of handshake sizes Should be removed before publishing In the following table, we compare the sizes of TLS 1.3- and AuthKEM-based handshakes. We give the transmission requirements for handshake authentication (public key + signature), and certificate chain (intermediate CA certificate public key and signature + root CA signature). For clarity, we are not listing post-quantum/traditional hybrid algorithms; we also omit mechanisms such as Certificate Transparency or OCSP stapling . We use ML-KEM-768 instead of the smaller ML-KEM-512 parameter set, as the former is currently used in experimental deployments. For signatures, we use Dilithium (standardized as ML-DSA), the "primary" algorithm selected by NIST for post-quantum signatures, as well as Falcon , the algorithm that offers smaller public key and signature sizes, but which NIST indicates can be used if the implementation requirements can be met. Size comparison of public-key cryptography in TLS 1.3 and AuthKEM handshakes.

Handshake	HS auth algorithm	HS Auth bytes	Certificate chain bytes	Sum
TLS 1.3	RSA-2048	528	784 (RSA-2048)	1312
TLS 1.3	Dilithium-2	3732	6152 (Dilithium-2)	9884
TLS 1.3	Falcon-512	1563	2229 (Falcon-512)	3792
TLS 1.3	Dilithium-2	3732	2229 (Falcon-512)	5961
AuthKEM	Kyber-768	2272	6152 (Dilithium-2)	8424
AuthKEM	Kyber-768	2272	2229 (Falcon-512)	4564

Note that although TLS 1.3 with Falcon-512 is the smallest instantiation, Falcon is very challenging to implement: signature generation requires (emulation of) 64-bit floating point operations in constant time. It is also very difficult to protect against other side-channel attacks, as there are no known methods of masking Falcon. In light of these difficulties, use of Falcon-512 in online handshake signatures may not be wise. Using AuthKEM with Falcon-512 in the certificate chain remains an attractive option, however: the certificate issuance process, because it is mostly offline, could perhaps be set up in a way to protect the Falcon implementation against attacks. Falcon signature verification is fast and does not require floating-point arithmetic. Avoiding online usage of Falcon in TLS 1.3 requires two implementations of the signature verification routines, i.e., ML-DSA and Falcon, on top of the key exchange algorithm. In all examples, the size of the certificate chain still dominates the TLS handshake, especially if Certificate Transparency SCT statements are included, which is relevant in the context of the WebPKI. However, we believe that if proposals to reduce transmission sizes of the certificate chain in the WebPKI context are implemented, the space savings of AuthKEM naturally become relatively larger and more significant. We discuss this in .

Related work

OPTLS This proposal draws inspiration from , which is in turn based on the OPTLS proposal for TLS 1.3 . However, these proposals require a non-interactive key exchange: they combine the client's public key with the server's long-term key. This imposes an extra requirement: the ephemeral and static keys MUST use the same algorithm, which this proposal does not require. Additionally, there are no post-quantum proposals for a non-interactive key exchange currently considered for standardization, while several KEMs are on the way.

Compressing certificates and certificate chains AuthKEM reduces the amount of data required for authentication in TLS. In recognition of the large increase in handshake size that a naive adoption of post-quantum signatures would affect, several proposals have been put forward that aim to reduce the size of certificates in the TLS handshake. proposes a certificate compression mechanism based on compression algorithms, but this is not very helpful to reduce the size of high-entropy public keys and signatures. Proposals that offer more significant reductions of sizes of certificate chains, such as , , , and all mainly rely on some form of out-of-band distribution of intermediate certificates or other trust anchors in a way that requires a robust update mechanism. This makes these proposals mainly suitable for the WebPKI setting; although this is also the setting that has the largest number of certificates due to the inclusion of SCT statements and OSCP staples . AuthKEM complements these approaches in the WebPKI setting. On its own the gains that AuthKEM offers may be modest compared to the large sizes of certificate chains. But when combined with compression or certificate suppression mechanisms such as those proposed in the referenced drafts, the reduction in handshake size when replacing Dilithium-2 by Kyber-768 for handshake authentication becomes significant again.

Organization After a brief introduction to KEMs, we will introduce the AuthKEM authentication mechanism. For clarity, we discuss unilateral and mutual authentication separately. In the remainder of the draft, we will discuss the necessary implementation mechanics, such as code points, extensions, new protocol messages and the new key schedule. The draft concludes with ah extensive discussion of relevant security considerations. A related mechanism for KEM-based PSK-style handshakes is discussed in .

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.

Terminology The following terms are used as they are in

client:

The endpoint initiating the TLS connection.

connection:

A transport-layer connection between two endpoints.

endpoint:

Either the client or server of the connection.

handshake:

An initial negotiation between client and server that establishes the parameters of their subsequent interactions within TLS.

peer:

An endpoint. When discussing a particular endpoint, "peer" refers to the endpoint that is not the primary subject of discussion.

receiver:

An endpoint that is receiving records.

sender:

An endpoint that is transmitting records.

server:

The endpoint that responded to the initiation of the TLS connection. i.e. the peer of the client.

Key Encapsulation Mechanisms As this proposal relies heavily on KEMs, which are not originally used by TLS, we will provide a brief overview of this primitive. Other cryptographic operations will be discussed later. A Key Encapsulation Mechanism (KEM) is a cryptographic primitive that defines the methods Encapsulate and Decapsulate. In this draft, we extend these operations with context separation strings, per HPKE :

Encapsulate(pkR, context_string):

Takes a public key, and produces a shared secret and encapsulation.

Decapsulate(enc, skR, context_string):

Takes the encapsulation and the private key. Returns the shared secret.

We implement these methods through the KEMs defined in to export shared secrets appropriate for using with the HKDF in TLS 1.3: Keys are generated and encoded for transmission following the conventions in . The values of context_string are defined in . Open question: Should we keep using HPKE, or just use "plain" KEMs, as in the original KEMTLS works? Please see the discussion at Issue #32.

Full 1.5-RTT AuthKEM Handshake Protocol Figure 1 below shows the basic KEM-authentication (AuthKEM) handshake, without client authentication: ServerHello ^ Key + key_share v Exch ^ --------> | {Finished} --------> | Auth [Application Data] --------> | <-------- {Finished} v [Application Data] <-------> [Application Data] + Indicates noteworthy extensions sent in the previously noted message. <> Indicates messages protected using keys derived from a [sender]_handshake_traffic_secret. {} Indicates messages protected using keys derived from a [sender]_authenticated_handshake_traffic_secret. [] Indicates messages protected using keys derived from [sender]_application_traffic_secret_N. Figure 1: Message Flow for KEM-Authentication (KEM-Auth) Handshake without client authentication. ]]> This basic handshake captures the core of AuthKEM. Instead of using a signature to authenticate the handshake, the client encapsulates a shared secret to the server's certificate public key. Only the server that holds the private key corresponding to the certificate public key can derive the same shared secret. This shared secret is mixed into the handshake's key schedule. The client does not have to wait for the server's Finished message before it can send data. The client knows that its message can only be decrypted if the server was able to derive the authentication shared secret encapsulated in the KEMEncapsulation message. Finished messages are sent as in TLS 1.3, and achieve full explicit authentication.

Client authentication For client authentication, the server sends the CertificateRequest message as in . This message can not be authenticated in the AuthKEM handshake: we will discuss the implications below. As in , section 4.4.2, if and only if the client receives CertificateRequest, it MUST send a Certificate message. If the client has no suitable certificate, it MUST send a Certificate message containing no certificates. If the server is satisfied with the provided certificate, it MUST send back a KEMEncapsulation message, containing the encapsulation to the client's certificate. The resulting shared secret is mixed into the key schedule. This ensures any messages sent using keys derived from it are covered by the authentication. The AuthKEM handshake with client authentication is given in Figure 2. ServerHello ^ Key + key_share v Exch ^ Server v Params ^ ^ | | {Certificate} --------> | Auth | <-------- {KEMEncapsulation} | Auth v {Finished} --------> | [Application Data] --------> | <------- {Finished} v [Application Data] <-------> [Application Data] + Indicates noteworthy extensions sent in the previously noted message. <> Indicates messages protected using keys derived from a [sender]_handshake_traffic_secret. {} Indicates messages protected using keys derived from a [sender]_authenticated_handshake_traffic_secret. [] Indicates messages protected using keys derived from [sender]_application_traffic_secret_N. Figure 2: Message Flow for KEM-Authentication (KEM-Auth) Handshake with client authentication. ]]> If the server is not satisfied with the client's certificates, it MAY, at its discretion, decide to continue or terminate the handshake. If it decides to continue, it MUST NOT send back a KEMEncapsulation message and the client and server MUST compute the encryption keys as in the server-only authenticated AuthKEM handshake. The Certificate remains included in the transcript. The client MUST NOT assume it has been authenticated. Unfortunately, AuthKEM client authentication requires an extra round-trip. Clients that know the server's long-term public KEM key MAY choose to use the abbreviated AuthKEM handshake and opportunistically send the client certificate as a 0-RTT-like message. This mechanism is discussed in .

Relevant handshake messages After the Key Exchange and Server Parameters phase of TLS 1.3 handshake, the client and server exchange implicitly authenticated messages. KEM-based authentication uses the same set of messages every time that certificate-based authentication is needed. Specifically:

• Certificate: The certificate of the endpoint and any per-certificate extensions. This message MUST be omitted by the client if the server did not send a CertificateRequest message (thus indicating that the client should not authenticate with a certificate). For AuthKEM, Certificate MUST include the long-term KEM public key. Certificates MUST be handled in accordance with , section 4.4.2.4.
• KEMEncapsulation: A key encapsulation against the certificate's long-term public key, which yields an implicitly authenticated shared secret.

Overview of key differences with RFC8446 TLS 1.3

• New types of signature_algorithms for KEMs.
• Public keys in certificates are KEM algorithms.
• New handshake message KEMEncapsulation.
• The key schedule mixes in the shared secrets from the authentication.
• The Certificate is sent encrypted with a new handshake encryption key.
• The client sends Finished before the server.
• The client sends data before the server has sent Finished.

Implicit and explicit authentication The data that the client MAY transmit to the server before having received the server's Finished is encrypted using ciphersuites chosen based on the client's and server's advertised preferences in the ClientHello and ServerHello messages. The ServerHello message can however not be authenticated before the Finished message from the server is verified. The full implications of this are discussed in the Security Considerations section. Upon receiving the client's authentication messages, the server responds with its Finished message, which achieves explicit authentication. Upon receiving the server's Finished message, the client achieves explicit authentication. Receiving this message retroactively confirms the server's cryptographic parameter choices.

Authenticating CertificateRequest The CertificateRequest message can not be authenticated during the AuthKEM handshake; only after the Finished message from the server has been processed, it can be proven as authentic. The security implications of this are discussed later. This is discussed in GitHub issue #16. We would welcome feedback there. Clients MAY choose to only accept post-handshake authentication. TODO: Should they indicate this? TLS Flag?

Implementation In this section we will discuss the implementation details such as extensions and key schedule.

Negotiation of AuthKEM Clients will indicate support for this mode by negotiating it as if it were a signature scheme (part of the signature_algorithms extension). We thus add these new signature scheme values (even though, they are not signature schemes) for the KEMs defined in Section 7.1. Note that we will be only using their internal KEM's API defined there. TBD, dhkem_p384_sha384 => TBD, dhkem_p521_sha512 => TBD, dhkem_x25519_sha256 => 0xFE01, dhkem_x448_sha512 => TBD, kem_x25519kyber768 => TBD, /*draft-westerbaan-cfrg-hpke-xyber768d00*/ } ]]> Please give feedback on which KEMs should be included When present in the signature_algorithms extension, these values indicate AuthKEM support with the specified key exchange mode. These values MUST NOT appear in signature_algorithms_cert, as this extension specifies the signing algorithms by which certificates are signed.

Protocol messages The handshake protocol is used to negotiate the security parameters of a connection, as in TLS 1.3. It uses the same messages, expect for the addition of a KEMEncapsulation message and does not use the CertificateVerify one. Protocol messages MUST be sent in the order defined in Section 4. A peer which receives a handshake message in an unexpected order MUST abort the handshake with a "unexpected_message" alert. The KEMEncapsulation message is defined as follows: opaque encapsulation<0..2^16-1>; } KEMEncapsulation; ]]> The encapsulation field is the result of a Encapsulate function. The Encapsulate() function will also result in a shared secret (ssS or ssC, depending on the peer) which is used to derive the AHS or MS secrets (See ). If the KEMEncapsulation message is sent by a server, the authentication algorithm MUST be one offered in the client's signature_algorithms extension. Otherwise, the server MUST terminate the handshake with an "unsupported_certificate" alert. If sent by a client, the authentication algorithm used in the signature MUST be one of those present in the supported_signature_algorithms field of the signature_algorithms extension in the CertificateRequest message. In addition, the authentication algorithm MUST be compatible with the key(s) in the sender's end-entity certificate. The receiver of a KEMEncapsulation message MUST perform the Decapsulate() operation by using the sent encapsulation and the private key of the public key advertised in the end-entity certificate sent. The Decapsulate() function will also result on a shared secret (ssS or ssC, depending on the Server or Client executing it respectively) which is used to derive the AHS or MS secrets. certificate_request_context is included to allow the recipient to identify the certificate against which the encapsulation was generated. It MUST be set to the value in the Certificate message to which the encapsulation was computed.

Cryptographic computations The AuthKEM handshake establishes three input secrets which are combined to create the actual working keying material, as detailed below. The key derivation process incorporates both the input secrets and the handshake transcript. Note that because the handshake transcript includes the random values from the Hello messages, any given handshake will have different traffic secrets, even if the same input secrets are used.

Key schedule for full AuthKEM handshakes AuthKEM uses the same HKDF-Extract and HKDF-Expand functions as defined by TLS 1.3, in turn defined by . Keys are derived from two input secrets using the HKDF-Extract and Derive-Secret functions. The general pattern for adding a new secret is to use HKDF-Extract with the Salt being the current secret state and the Input Keying Material (IKM) being the new secret to be added. The notable differences are:

• The addition of the Authenticated Handshake Secret and a new set of handshake traffic encryption keys.
• The inclusion of the SSs and SSc (if present) shared secrets as IKM to Authenticated Handshake Secret and Main Secret, respectively.

The full key schedule proceeds as follows: HKDF-Extract = Early Secret | +--> Derive-Secret(., "ext binder" | "res binder", "") | = binder_key | +--> Derive-Secret(., "c e traffic", ClientHello) | = client_early_traffic_secret | +--> Derive-Secret(., "e exp master", ClientHello) | = early_exporter_master_secret v Derive-Secret(., "derived", "") | v (EC)DHE -> HKDF-Extract = Handshake Secret | +--> Derive-Secret(., "c hs traffic", | ClientHello...ServerHello) | = client_handshake_traffic_secret | +--> Derive-Secret(., "s hs traffic", | ClientHello...ServerHello) | = server_handshake_traffic_secret v Derive-Secret(., "derived", "") = dHS | v SSs -> HKDF-Extract = Authenticated Handshake Secret | +--> Derive-Secret(., "c ahs traffic", | ClientHello...KEMEncapsulation) | = client_handshake_authenticated_traffic_secret | +--> Derive-Secret(., "s ahs traffic", | ClientHello...KEMEncapsulation) | = server_handshake_authenticated_traffic_secret v Derive-Secret(., "derived", "") = dAHS | v SSc||0 * -> HKDF-Extract = Main Secret | +--> Derive-Secret(., "c ap traffic", | ClientHello...client Finished) | = client_application_traffic_secret_0 | +--> Derive-Secret(., "s ap traffic", | ClientHello...server Finished) | = server_application_traffic_secret_0 | +--> Derive-Secret(., "exp master", | ClientHello...server Finished) | = exporter_master_secret | +--> Derive-Secret(., "res master", ClientHello...server Finished) = resumption_master_secret *: if client authentication was requested, the `SSc` value should be used. Otherwise, the `0` value is used. ]]>

Computations of KEM shared secrets The operations to compute SSs or SSc from the client are: The operations to compute SSs or SSc from the server are:

Explicit Authentication Messages AuthKEM upgrades implicit to explicit authentication through the Finished message. Note that in the full handshake, AuthKEM achieves explicit authentication only when the server sends the final Finished message (the client is only implicitly authenticated when they send their Finished message). Full downgrade resilience and forward secrecy is achieved once the AuthKEM handshake completes. The key used to compute the Finished message MUST be computed from the MainSecret using HKDF (instead of a key derived from HS as in ). Specifically: The verify_data value is computed as follows. Note that instead of what is specified in , we use the full transcript for both server and client Finished messages: Any records following a Finished message MUST be encrypted under the appropriate application traffic key as described in . In particular, this includes any alerts sent by the server in response to client Certificate and KEMEncapsulation messages. See for a full treatment of implicit and explicit authentication.

Security Considerations

Implicit authentication Because preserving a 1/1.5RTT handshake in KEM-Auth requires the client to send its request in the same flight when the ServerHello message is received, it can not yet have explicitly authenticated the server. However, through the inclusion of the key encapsulated to the server's long-term secret, only an authentic server should be able to decrypt these messages. However, the client can not have received confirmation that the server's choices for symmetric encryption, as specified in the ServerHello message, were authentic. These are not authenticated until the Finished message from the server arrived. This may allow an adversary to downgrade the symmetric algorithms, but only to what the client is willing to accept. If such an attack occurs, the handshake will also never successfully complete and no data can be sent back. If the client trusts the symmetric algorithms advertised in its ClientHello message, this should not be a concern. A client MUST NOT accept any cryptographic parameters it does not include in its own ClientHello message. If client authentication is used, explicit authentication is reached before any application data, on either client or server side, is transmitted. Application Data MUST NOT be sent prior to sending the Finished message, except as specified in Section 2.3 of . Note that while the client MAY send Application Data prior to receiving the server's last explicit Authentication message, any data sent at that point is, being sent to an implicitly authenticated peer.

Authentication of Certificate Request Due to the implicit authentication of the server's messages during the full AuthKEM handshake, the CertificateRequest message can not be authenticated before the client received Finished. The key schedule guarantees that the server can not read the client's certificate message (as discussed above). An active adversary that maliciously inserts a CertificateRequest message will also result in a mismatch in transcript hashes, which will cause the handshake to fail. However, there may be side effects. The adversary might learn that the client has a certificate by observing the length of the messages sent. There may also be side effects, especially in situations where the client is prompted to e.g. approve use or unlock a certificate stored encrypted or on a smart card.

Other security considerations

• Because the Main Secret is derived from both the ephemeral key exchange, as well as from the key exchanges completed for server and (optionally) client authentication, the MS secret always reflects the peers' views of the authentication status correctly. This is an improvement over TLS 1.3 for client authentication.
• The academic works proposing AuthKEM (KEMTLS) contains an in-depth technical discussion of and a proof of the security of the handshake protocol without client authentication , .
• The work proposing the variant protocol , with pre-distributed public keys (the abbreviated AuthKEM handshake) has a proof for both unilaterally and mutually authenticated handshakes.
• We have proofs of the security of KEMTLS and KEMTLS-PDK in Tamarin.
• Application Data sent prior to receiving the server's last explicit authentication message (the Finished message) can be subject to a client certificate suite downgrade attack. Full downgrade resilience and forward secrecy is achieved once the handshake completes.
• The client's certificate is kept secret from active observers by the derivation of the client_authenticated_handshake_secret, which ensures that only the intended server can read the client's identity.
• If AuthKEM client authentication is used, the resulting shared secret is included in the key schedule. This ensures that both peers have a consistent view of the authentication status, unlike .

References Normative References 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 updates RFCs 5705 and 6066, and obsoletes RFCs 5077, 5246, and 6961. This document also specifies new requirements for TLS 1.2 implementations. This document describes a scheme for hybrid public key encryption (HPKE). This scheme provides a variant of public key encryption of arbitrary-sized plaintexts for a recipient public key. It also includes three authenticated variants, including one that authenticates possession of a pre-shared key and two optional ones that authenticate possession of a key encapsulation mechanism (KEM) private key. HPKE works for any combination of an asymmetric KEM, key derivation function (KDF), and authenticated encryption with additional data (AEAD) encryption function. Some authenticated variants may not be supported by all KEMs. We provide instantiations of the scheme using widely used and efficient primitives, such as Elliptic Curve Diffie-Hellman (ECDH) key agreement, HMAC-based key derivation function (HKDF), and SHA2. This document is a product of the Crypto Forum Research Group (CFRG) in the IRTF. sn3rd AWS AWS Cloudflare Module-Lattice-Based Key-Encapsulation Mechanism (ML-KEM) is a quantum-resistant key-encapsulation mechanism (KEM). This document specifies algorithm identifiers and ASN.1 encoding format for ML-KEM in public key certificates. The encoding for public and private keys are also provided. [EDNOTE: This document is not expected to be finalized before the NIST PQC Project has standardized PQ algorithms. This specification will use object identifiers for the new algorithms that are assigned by NIST, and will use placeholders until these are released.] 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. 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. Informative References This document specifies a simple Hashed Message Authentication Code (HMAC)-based key derivation function (HKDF), which can be used as a building block in various protocols and applications. The key derivation function (KDF) is intended to support a wide range of applications and requirements, and is conservative in its use of cryptographic hash functions. This document is not an Internet Standards Track specification; it is published for informational purposes. University of Waterloo Radboud University and Max Planck Institute for Security and Privacy Radboud University University of Waterloo Radboud University and Max Planck Institute for Security and Privacy Radboud University Brave Software Cloudflare, Inc. University of Waterloo Radboud University Radboud University National Institute for Standards and Technology IBM Research ENS, CNRS, INRIA and Columbia University Radboud University n.d. n.d. This document specifies algorithm identifiers and ASN.1 encoding formats for elliptic curve constructs using the curve25519 and curve448 curves. The signature algorithms covered are Ed25519 and Ed448. The key agreement algorithms covered are X25519 and X448. The encoding for public key, private key, and Edwards-curve Digital Signature Algorithm (EdDSA) structures is provided. Cisco Facebook Cloudflare Mozilla The organizational separation between operators of TLS and DTLS endpoints and the certification authority can create limitations. For example, the lifetime of certificates, how they may be used, and the algorithms they support are ultimately determined by the Certification Authority (CA). This document describes a mechanism to overcome some of these limitations by enabling operators to delegate their own credentials for use in TLS and DTLS without breaking compatibility with peers that do not support this specification. Cloudflare Cloudflare This memo defines X25519Kyber768Draft00, a hybrid post-quantum KEM, for HPKE (RFC9180). This KEM does not support the authenticated modes of HPKE. MPI-SP & Radboud University Cloudflare This memo specifies a preliminary version ("draft00", "v3.02") of Kyber, an IND-CCA2 secure Key Encapsulation Method. About This Document This note is to be removed before publishing as an RFC. The latest revision of this draft can be found at https://bwesterb.github.io/draft-schwabe-cfrg-kyber/draft-cfrg- schwabe-kyber.html. Status information for this document may be found at https://datatracker.ietf.org/doc/draft-cfrg-schwabe-kyber/. Source for this draft and an issue tracker can be found at https://github.com/bwesterb/draft-schwabe-cfrg-kyber. PQShield Brave Software Radboud University and MPI-SP University of Waterloo This document gives a construction in which (long-term) KEM public keys are used in the place of TLS PSK keys, avoiding concerns that may affect systems that use symmetric-key-based PSK, such as requiring key diversification and protection of symmetric-keys' confidentiality. This mechanism is inspired by AuthKEM (and could use AuthKEM certificate public keys for resumption), but can be independently implemented. 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. 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. Mozilla Cloudflare Apple Inc. TLS 1.3 [RFC8446] specifies a signed Diffie-Hellman exchange modelled after SIGMA [SIGMA]. This design is suitable for endpoints whose certified credential is a signing key, which is the common situation for current TLS servers. This document describes a mode of TLS 1.3 in which one or both endpoints have a certified DH key which is used to authenticate the exchange. Note to Readers Source for this draft and an issue tracker can be found at https://github.com/ekr/draft-rescorla-tls13-semistatic-dh (https://github.com/ekr/draft-rescorla-tls13-semistatic-dh). In TLS handshakes, certificate chains often take up the majority of the bytes transmitted. This document describes how certificate chains can be compressed to reduce the amount of data transmitted and avoid some round trips. Mozilla This draft defines a new TLS Certificate Compression scheme which uses a shared dictionary of root and intermediate WebPKI certificates. The scheme smooths the transition to post-quantum certificates by eliminating the root and intermediate certificates from the TLS certificate chain without impacting trust negotiation. It also delivers better compression than alternative proposals whilst ensuring fair treatment for both CAs and website operators. It may also be useful in other applications which store certificate chains, e.g. Certificate Transparency logs. Windy Hill Systems, LLC Cisco Meta Platforms, Inc. This document specifies a "compact" version of TLS 1.3 and DTLS 1.3. It saves bandwidth by trimming obsolete material, tighter encoding, a template-based specialization technique, and alternative cryptographic techniques. cTLS is not directly interoperable with TLS 1.3 or DTLS 1.3 since the over-the-wire framing is different. A single server can, however, offer cTLS alongside TLS or DTLS. AWS AWS Cloudflare Mozilla A TLS client or server that has access to the complete set of published intermediate certificates can inform its peer to avoid sending certificate authority certificates, thus reducing the size of the TLS handshake. Google LLC Google LLC Cloudflare This document describes Merkle Tree certificates, a new certificate type for use with TLS. A relying party that regularly fetches information from a transparency service can use this certificate type as a size optimization over more conventional mechanisms with post- quantum signatures. Merkle Tree certificates integrate the roles of X.509 and Certificate Transparency, achieving comparable security properties with a smaller message size, at the cost of more limited applicability.

Open points of discussion The following are open points for discussion. The corresponding Github issues will be linked.

Authentication concerns for client authentication requests. Tracked by Issue #16. The certificate request message from the server can not be authenticated by the AuthKEM mechanism. This is already somewhat discussed above and under security considerations. We might want to allow clients to refuse client auth for scenarios where this is a concern.

Interaction with signing certificates Tracked by Issue #20. In the current state of the draft, we have not yet discussed combining traditional signature-based authentication with KEM-based authentication. One might imagine that the Client has a sigining certificate and the server has a KEM public key. In the current draft, clients MUST use a KEM certificate algorithm if the server negotiated AuthKEM.

Acknowledgements Early versions of this work were supported by the European Research Council through Starting Grant No. 805031 (EPOQUE). Part of this work was supported by the NLNet NGI Assure theme fund project "Standardizing KEMTLS"