[lamps] Re: Composite Signatures and KEM open issues that need feedback

[Mike Ounsworth <[e-mail]>]

Thanks Tim.

 

Bit swamped at the moment with feedback and summer vacations, but I parked it here so we don’t lose it:

https://github.com/lamps-wg/draft-composite-sigs/issues/32

https://github.com/lamps-wg/draft-composite-kem/issues/62

 

---

Mike Ounsworth

 

From: Tim Hollebeek <tim.hollebeek=[e-mail]> 
Sent: Friday, August 16, 2024 12:20 PM
To: John Gray <John.Gray=[e-mail]>; [e-mail]
Subject: [EXTERNAL] [lamps] Re: Composite Signatures and KEM open issues that need feedback

 

 

DigiCert's CA engineering team has some comments on the open issues related

to the composite-sigs draft. We're going to put them in one email just

because we have comments on quite a few of them.

 

ISSUE #1

(Github issue:  https://github.com/lamps-wg/draft-composite-sigs/issues/9)

 

ASN.1 wrapping confuses people. This came up in the hash-based signatures

updates last call. Nobody knows what ASN.1 is, or what the consequences

of omiting it are (to be clear, there are really none).

 

We agree that this is largely a question of people being unfamiliar with

ASN.1, and that explanatory text is sufficient. All that is needed is a

clear explanation of example what the BIT STRING is, and explaining that

it's simply the bits of the key itself seems pretty straightforward.

 

One can then say something similar to: 

 

"In some situations and protocols, the key might be wrapped in ASN.1 or 

may have some other additional decoration or encoding. If so, such wrapping 

MUST be removed prior to encoding the key itself as a BIT STRING."

 

Hopefully that makes things crystal clear.

 

ISSUE #2

(Github Issue: https://github.com/lamps-wg/draft-composite-sigs/issues/19)

 

We don't think it's worth the extra complexity and expense of an additional

hash operation just to achieve a fixed size output. The variation in size

is already pretty small.

 

ISSUE #3

(Github issue: https://github.com/lamps-wg/draft-composite-sigs/issues/6)

 

Again, we don't believe the additional complexity is worth it for a pretty

trivial improvement in the private key size. But it's not a strong opinion,

we could go either way.

 

Open Issues affect both Composite Signatures and Composite KEM:

ISSUE #4

 

Chair hat off, I and the CA team are concerned about the slow progress of 

the composite signature work. In particular, tying it to the Composite KEM

draft and waiting for the CFRG work on KEM Combiners seems like an

absolutely horrible idea to us. We would like to see Composite Signatures

progress ASAP.

 

ISSUE #5

(Github issues:

https://github.com/lamps-wg/draft-composite-kem/issues/37

https://github.com/lamps-wg/draft-composite-sigs/issues/24

https://github.com/lamps-wg/draft-composite-sigs/issues/23)

 

This is a fun one, and we've spent quite a bit of time discussing it

internally.

 

In particular, we're still debating the question about exactly how many

backwards compatibility options are really necessary. For example, given

that you already need to add lattice, is it really necessary to allow

PKCS15 to continue to exist? For RSA, there's the reasonable argument that

that might be all you have in your validated hardware/software, but if you

have RSA as a primitive, can't you do PSS instead of PKCS15? Remember, you

already have to make changes on both the signing and verify side anyway.

 

We're trending in the direction of thinking that the primary decision is

the security level and post-quantum algorithm, and the classical side is

just determined by what "makes sense" for that security level and algorithm.

 

So what you really want is something like "id-SL1-MLKEM-RSA" where the

document specifies exactly what "RSA" means in the context of a SL1 MLKEM512

composite, e.g. RSA4096-PSS-SHA512. This basically means striving for

at most one combination for each triple of (security level, PQC flavor, classical 

flavor) and eliminating unnecessary complexity and diversity of options in what 

is essentially a redundancy/backup mechanism.

 

The basic idea is to more aggressively standardize the backwards 

compatibility options to only what's actually necessary, instead of

trying to be backwards compatible with the universe of current behavior,

which both unnecessarily complicates things, and preserves some practices

(e.g. PKCS15) longer than is perhaps prudent.

 

-Tim

 

From: John Gray <John.Gray=[e-mail] <mailto:John.Gray=[e-mail]> > 
Sent: Thursday, August 1, 2024 5:41 PM
To: [e-mail] <mailto:[e-mail]> 
Subject: [lamps] Composite Signatures and KEM open issues that need feedback

 

Hello Lamps.  

Thanks for the feedback at IETF 120 for composite signatures and composite KEMs.   We the authors have compiled together all the currently open questions about composites into one email (sorry it is so long).

Feedback on-list is great. Discussion directly on the linked github issue is better.

If you’re going to comment on the mailing list. Please carefully tag which issue you are discussing so that it stays somewhat orderly … “Responding to ISSUE #3”.

 

Open Questions on Composite Signatures:

ISSUE #1

(Github issue:  https://github.com/lamps-wg/draft-composite-sigs/issues/9)

 

The ML-DSA public key should be an unwrapped BIT STRING with no ASN.1 type around them.   Currently the ML-DSA draft draft-ietf-lamps-dilithium-certificates uses this:

pk-MLDSA PUBLIC-KEY ::= {

      IDENTIFIER id-MLDSA

       -- KEY no ASN.1 wrapping --

       PARAMS ARE absent

       CERT-KEY-USAGE { nonRepudiation, digitalSignature,

                       keyCertSign, cRLSign }

       --- PRIVATE-KEY no ASN.1 wrapping --

     }

We could try using something like an ENCODED BY id-rawkey as in:

   id-raw-key ::= SOME OBJECT IDENTIFIER

   pk-CompositeSignature {OBJECT IDENTIFIER:id,

     FirstPublicKeyType,SecondPublicKeyType }

       PUBLIC-KEY ::= {

         IDENTIFIER id

         KEY SEQUENCE {

           firstPublicKey BIT STRING (CONTAINING FirstPublicKeyType  | ENCODED BY id-raw-key),

           secondPublicKey BIT STRING (CONTAINING SecondPublicKeyType | ENCODED BY id-raw-key)

         }

         PARAMS ARE absent

         CERT-KEY-USAGE { digitalSignature, nonRepudiation, keyCertSign, cRLSign}

       }

Or just have some text that explains the BIT STRING is a raw key without the extra ASN.1 type wrapping.  Does the working group have a preference on this matter?   The authors think adding some explanatory text should be sufficient.

 

ISSUE #2

(Github Issue: https://github.com/lamps-wg/draft-composite-sigs/issues/19)

 

Do we make the Domain separator Hash (DER (OID)) instead of just DER(OID)?  The one advantage is we end up with a fixed length Domain separator.  The authors don’t think this is required, but are willing to make the change if the working group would like to see this done.

 

ISSUE #3

(Github issue: https://github.com/lamps-wg/draft-composite-sigs/issues/6)

Should we consider compacting the CompositeSignaturePrivateKey format?  For example, today it is:

CompositeSignaturePrivateKey ::= SEQUENCE SIZE (2) OF OneAsymmetricKey

We could compact it to:

CompositeSignaturePrivateKey ::= SEQUENCE SIZE (2) OF OCTET STRING

Then implementations would need to recompose the OneAsymmetricKey  using the combination of settings give by the OID.  This removes the redundant information from the CompositeSignaturePrivateKey because it is now carried in the OID representation itself.

The authors don’t have  strong opinion on whether this should be changed.  The two benefits:

1.      Smaller private keys (maybe a couple percent)

2.      Aligns with the compact format used in the public key.

3.	It makes it a bit more difficult for implementors (but not too much).  

---------

Open Issues affect both Composite Signatures and Composite KEM:

ISSUE #4

Timing. Does LAMPS have an official or unofficial milestone for publishing?  Answer could be different for KEMs and Sigs. Obviously, there’s going to be a flurry of PQC drafts from LAMPS and others once FIPS 203, 204, 205 are out. Are we trying to get these into that wave, or not? In particular, should we wait X years for CFRG to finish their KEM Combiners activity, or should we publish this and fix it later if we need to? (The authors and the general feel at LAMPS 120 was to publish-now-fix-later).

 

ISSUE #5

(Github issues:  

 <https://github.com/lamps-wg/draft-composite-kem/issues/37> https://github.com/lamps-wg/draft-composite-kem/issues/37

 <https://github.com/lamps-wg/draft-composite-sigs/issues/24> https://github.com/lamps-wg/draft-composite-sigs/issues/24 

https://github.com/lamps-wg/draft-composite-sigs/issues/23

Should the RSA key size be specified by the OID, or left free?
Related sub question: if specified, which RSA key sizes should we support? Currently we have {2048, 3072}, but we’ve been asked by an implementer to add 4096. Should it then be {2048, 4096}? Jan is advocating for having all three RSA sizes.

 

Arguments for removing the key size restriction: it completely avoids that related sub question.


Arguments against:  we probably should provide guidance on how to pair RSA key sizes with the ML-KEM and KDF parameter.  Removing the RSA key size does not shorten the current list; you need a minimum of 4 RSA combos to hit: MLDSA44+PSS, MLDSA44+PKCS1, MLDSA65+PSS, MLDSA65+PKCS1, and the equivalents on the KEM side.  Reminder: having both L1/2, and L3 with RSA is sort of about matching security levels (although we would be quite happy to call all levels of RSA “NIST PQC L1”), but it’s more about people trying to add PQ to an RSA deployment and they only want to implement one size of ML-*. Maybe this is a weak argument? But then if we only go with one PQ level, then which ML-DSA and which ML-KEM is “the one”?    

 

For signature the authors are intending to add two new combinations with RSA 4096.  We suggest:

 

id-MLDSA65-RSA4096-PSS-SHA512

id-MLDSA65-RSA4096-PKCS15-SHA512 

 

For KEM we are intended to add 

id-MLKEM512-RSA4096 

-----------

Open Issues affecting Composite KEM

ISSUE #6

(Github:  https://github.com/lamps-wg/draft-composite-kem/issues/40)

 

Combiner construction. We’re on-path to align with OpenPGP and X-Wing. We greatly appreciate the interaction with Quynh Dang to make sure this is FIPS-compliant.

Sorry for the confusing notation on the slides and in draft-04.

The intention is to get as close to the OpenPGP construction as we can, which is:

SHA3-*( mlkemSS || tradSS || tradCT || tradPK || domSep )


Note: X-wing puts its spaceship ascii art domain separator first, which we believe will not pass SP800.56Cr2; We believe it has to be at the end to fit 56Cr2’s FixedInfo.

The authors will work with the authors of draft-openpgp-pqc, X-Wing, and the forthcoming KEM Combiners draft at CFRG to align hopefully to the point of binary compatibility. I don’t think we actually need any WG feedback, unless there are objections.

 <https://github.com/lamps-wg/draft-composite-kem/issues/40> https://github.com/lamps-wg/draft-composite-kem/issues/40
https://github.com/lamps-wg/draft-composite-kem/issues/45

 

ISSUE #7

(Github: https://github.com/lamps-wg/draft-composite-kem/issues/54)

 

For a security proof of the ML-KEM + ECDH combos, we can point to the X-Wing paper. We should have a similar proof for the RSA-OAEP combos. Not sure how to go about attracting someone to help with this, or if we should attempt proof-writing ourselves.

 

ISSUE #8

(Github:  https://github.com/lamps-wg/draft-composite-kem/issues/52)

KEM domain separators.
Currently we are using DER(OID) as the domain separator. We think it is desirable that a shared secret derived for CMS cannot be swapped into, for example, an OpenPGP or HPKE context. But maybe that does not need to be handled at the KEM algorithm level, maybe it is already handled at the protocol level (CMS KEMRI, for example), and so we should align on domain separators with OpenPGP?

For X-Wing in particular, for that one we could take the “\.//^\”, but as mentioned above, that wouldn’t give binary compatability anyway since they put it at the beginning, and we have to put it at the end for FIPS reasons.  Do we want binary compatibility with OpenPGP KEM and XWing.  If this is true then shouldn’t we have one composite KEM primitive draft, and then drafts that specify usage in OpenPGP and CMS.  We also need to think about the encoding of the public key.

 

ISSUE #9

(Github: https://github.com/lamps-wg/draft-composite-kem/issues/48)

The term “DHKEM” and “RSAOAEPKEM”. Are those registered and already used? In particular, does RFC 9180 have a monopoly on the term “DHKEM”? If so, the authors are happy to take suggestions for what to call our abstract pseudocode in 2.3.3.
 <https://github.com/lamps-wg/draft-composite-kem/issues/48> https://github.com/lamps-wg/draft-composite-kem/issues/48.   Are we worried about name collisions of Pseudo code in RFC’s.   What do we rename it to?

 

ISSUE #10

(Github: https://github.com/lamps-wg/draft-composite-kem/issues/50

https://github.com/lamps-wg/draft-composite-kem/issues/49)

 

Binding public keys in the KDF. Basically, CMS is used by a lot of embedded and hardware things – a smartcard applet that decrypts S/MIME is an example here. It’s never been true in the past that the device needs the public key in order to do a decryption. Are we asking for trouble if we suddenly make that a requirement of the RSA / ECC side of the hybrid?
Note that we only need it at the combiner level, not inside the RSA / ECC decryption routine, which remains unmodified.
The authors vote is that this is an ok thing to prescribe, but: 1) We are not embedded hardware vendors, and 2) we’ve been asking this question for a while and have received zero community feedback on it.   If we think this is needed in composite, it will have to added.   As an example, in  <https://docs.oracle.com/javacard/3.0.5/api/javacard/security/ECPrivateKey.html> https://docs.oracle.com/javacard/3.0.5/api/javacard/security/ECPrivateKey.html  does not contain the public key.

 

ISSUE #11

(Github: https://github.com/lamps-wg/draft-composite-kem/issues/51)

 

KDF = SHA3 … what about SHA2.
This question should probably have the same answer as Dan van Geest’s related question about draft-cms-kyber.
The argument for is that even though ML-KEM needs SHA3 internally, this may not always be available to the layer of code doing the combiner, or the CMS EnvelopedData.  Just a note, because the KDF is part of the composite OID, adding HKDF-SHA2 will 2x the size of the list. If the WG wants us to do this, great, but please don’t complain at us about the size of the list.

 

The authors suggest, rather than 2x'ing the whole list, we do the following:

 

All RSA combinations use HKDF-SHA2.

Each of the P256 and brainpoolP256 combinations are offered with both SHA3 (to align with X-Wing), and HKDF-SHA2. The new list would then be:

 

| Composite KEM                     | KDF       |

|---------                          | --------  |

| id-MLKEM512-ECDH-P256             | SHA3-256  |

| id-MLKEM512-ECDH-P256             | HKDF-SHA2 |

| id-MLKEM512-ECDH-brainpoolP256r1  | SHA3-256  |

| id-MLKEM512-ECDH-brainpoolP256r1  | HKDF-SHA2 |

| id-MLKEM512-X25519                | SHA3-256  |

| id-MLKEM512-RSA2048               | HKDF-SHA2 |

| id-MLKEM512-RSA3072               | HKDF-SHA2 |

| id-MLKEM512-RSA4096            | HKDF-SHA2 |

| id-MLKEM768-ECDH-P256             | SHA3-384  |

| id-MLKEM768-ECDH-P256             | HKDF-SHA2 |

| id-MLKEM768-ECDH-brainpoolP256r1  | SHA3-384  |

| id-MLKEM768-ECDH-brainpoolP256r1  | HKDF-SHA2 |

| id-MLKEM768-X25519                | SHA3-384  |

| id-MLKEM1024-ECDH-P384            | SHA3-512  |

| id-MLKEM1024-ECDH-brainpoolP384r1 | SHA3-512  |

| id-MLKEM1024-X448                 | SHA3-512  |

{: #tab-kem-algs title="Composite KEM key types"}

 

ISSUE #12

(Github: https://github.com/lamps-wg/draft-composite-kem/issues/47)

Should we remove ML-KEM512 and only offer 768 and 1024?  

 

Thanks in advance for all your feedback to move these drafts forward,

 

The Composite Signature and KEM authors

 

Any email and files/attachments transmitted with it are intended solely for the use of the individual or entity to whom they are addressed. If this message has been sent to you in error, you must not copy, distribute or disclose of the information it contains. Please notify Entrust immediately and delete the message from your system.