IKE and IKEv2 Authentication Using the Elliptic Curve Digital Signature Algorithm (ECDSA)
draft-ietf-ipsec-ike-auth-ecdsa-06

[Note]: 'An updated draft was submitted since this was on the agenda.  It contains several minor corrections; the biggest is to assign values for IKEv2 as well as IKE.' added by Russ Housley

[Note]: 'Draft-ietf-ipsec-ike-auth-ecdsa-06.txt in now available.  It contains several minor corrections; the biggest is to assign values for IKEv2 as well as IKE.' added by Russ Housley

[Note]: 'There were no objections to approving this document at the 7/06/2006 IESG Teleconference.  It will go into "Approved" when it finishes Last Call.' added by Amy Vezza

[Ballot comment]
Nits from Gen-ART review by Gonzalo Camarillo:

The boilerplates should be updated to be compliant with current IETF policies.

Acronyms in the title of the draft should be expanded.

In section 3, the draft says: The document defines the following... the phrase would be clearer if it said instead: *This* document defines...

The second paragraph of page 4 refers the IANA registry by giving a pointer to a reference ([IANA]). The reference [IANA] points to the registry using its URI. IANA registries should not be referenced by their URIs; they should be referenced by their names. In this case, "The IANA Registry for IPSEC Authentication Methods". This should be changed in both Sections 3 and 5.

The Security Considerations and IANA Considerations Sections are typically the last sections before the References and the Authors' Addresses. This draft has a section (Section 6) after the IANA Considerations section. I suggest that that section is converted into an appendix or moved before the Security Considerations Section.

Nits from Last Call review by Elwyn Davies:

s1, para 2 contains the assertion 'For any given level of security against the best attacks known, ECDSA signatures are smaller than RSA signatures and ECDSA keys require less bandwidth than DSA keys; there are also advantages of computational speed and efficiency in many settings.'  If the reference [LV] backs up this assertion it would be good to reference it here also.  If not it would be good to have some other reference that justifies this assertion.

(actually editorial+): s5: It would be useful to give the three methods mnemonic names which can be placed in the IANA registry and document them in s5.


s1: Acronyms IPsec, RSA and DSA need expansion.

s3: Acronym MAC needs expansion.

s3, table: To avoid confusion with references, the code points (e.g., [19]) currently enclosed in square brackets would be better in round brackets.

s6: It might be helpful to explicitly bracket y^2 = x^3 - 3 x + b  modulo p  - I think this is supposed to be y^2 = (x^3 - 3 x + b)  modulo p

s7.1: BTW IKEv2 has made it to RFC - RFC4306.

IANA comments

Upon approval of this document the IANA will assign the following 3 new
authentication algorithms:


          Digital        Diffie-
IANA    Signature  Hellman            Hash
Value  Algorithm  Group              Function
----- --------- ------------ -------------
9          ECDSA        P-256 [19]      SHA-256 [4]
10        ECDSA        P-384 [20]      SHA-384 [5]
11        ECDSA        P-521 [21]      SHA-512 [6]


These registrations will take place at the following location:
http://www.iana.org/assignments/ipsec-registry

It is not clear in which registry at the above location these
registrations will go in. Further clarification is needed.

Posted related IPR disclosure: Certicom's Statement about IPR Claimed in RFC 3446, RFC 2409, draft-ietf-tls-ecc-12, draft-ietf-ipsec-ike-auth-ecdsa-05, and draft-ietf-ipsec-ike-ecp-groups-02 and other IETF specifications using ECC technology