The RC4-HMAC Kerberos Encryption Types Used by Microsoft Windows
draft-jaganathan-rc4-hmac-03

[Ballot comment]
I am also concerned about the "export language" raised by others, but
  I see no reason to "pile on" another DISCUSS on that point.  I'm sure
  it will be fixed.

  In several places, SGN_ALG contains an integrity algorithm indicator.
  And the document says that "11 00 - HMAC".  I hope that this means
  HMAC-MD5 and that HMAC-SHA1 might be used in the future with a value
  other than "11 00."

[Ballot discuss]
Section says:
  >
  > The pseudo-random operation [RFC3961] for both enctypes above is
  > defined as follows:
  >
  >        pseudo-random(K, S) = HMAC-SHA1(K, S)
  >
  So, a normative reference to SHA-1 is needed.

[Ballot discuss]
As has been pointed out by Cullen and Jari, the export control
language in this draft is inaccurate.  Please revise the claim about
3DES being unavailable for export to claim that when Windows 2000 was
released, 3DES was unavailable.

Also, the IANA actions section is inaccurate; the references for the
rc4 enctypes should be updated to point to this specification.

[Ballot comment]
Like Cullen, I am concerned about the export control
statements. We know for a fact that strong crypto
is exportable, just look at what a random U.S. based
security product vendor sells elsewhere in the world;
of course its possible that there are conditions
relating to the type of product or how its delivered,
what country we are talking about etc. But it serves
the IETF no good to make make broad statements
like this in RFCs, someone may actually believe
them and not use the best possible crypto. Please
remove or qualify.

[Note]: 'The expert has reviewed the draft and sent comments to the author.  I believe a revised ID is required to address these comments.' added by Sam Hartman

Hi.  I've taken a brief look at the draft.  You are missing an IANA
section, which should request the registration of the enctypes.

I am asking Ken Raeburn as the RFC 3961 IANA expert to take a look at
this document.


Also, you should probably update the security considerations section
to mention related key attacks against RC4; see
draft-harris-ssh-arcfour-fixes for a recent security consideration
section for an RC4 system approved by the IESG.