From: Leslie Daigle
To: Dean Anderson
Cc: iab; iesg
Subject: Appeal of IESG Decision of July 10, 2006

On September 10, the IAB received an appeal from Dean Anderson on the
IESG’s decision to Last Call draft-ietf-grow-anycast-03.

1. SUMMARY OF IAB RESPONSE

The appeal is denied.

2. BACKGROUND

draft-ietf-grow-anycast, “Operation of Anycast Services” is a work item
of the GROW WG. In response to a request [0] from the WG chair, Geoff
Huston, to advance the document to BCP the IESG issued an IETF Last Call
on June 2, 2006. Mr. Anderson appealed the issuing of the Last Call to the
IESG which denied the appeal. Mr. Anderson subsequently appealed to the IAB.

Timeline:
Jun 2, 2006: The IESG issues an IETF Last Call for draft-ietf-grow-anycast-03.
Jun 13, 2006: Mr. Anderson appeals the Last Call.
Jul 1, 2006: Mr. Huston issues a WG LC for draft-ietf-grow-anycast-04
Jul 10, 2006: The IESG rejects Mr. Anderson’s appeal.
Jul 26, 2006: Mr. Huston calls WG consensus.
Aug 30, 2006: The IESG issues an IETF Last Call for draft-ietf-grow-anycast-04.
Sep 10, 2006: Mr. Anderson appeals to the IAB.

Mr. Anderson’s appeal raises four objections, one substantive and three
procedural:

S1. The document is technically flawed.
P1. The WG Chair (Mr. Huston) was incorrect to declare consensus
on the document as advanced and therefore the IESG was incorrect
to Last Call it.
P2. The document relies upon claims by Daniel Karrenberg about the
behavior of Anycast DNS and these claims were scientifically
fraudulent.
P3. That a number of IETF participants, including, but not limited to
IESG members Brian Carpenter and David Kessens had conflicts
of interest which should have caused them to recuse themselves
from decisions.

3. ANALYSIS

S1. Technical Flaws
Because the disposition of this document has not yet been decided on
by the IESG and part of the purpose of IESG review is evaluating the
quality of the technology, the possible existence of technical flaws
is not yet appealable. The IAB notes that the IETF Last Call has expired
and encourages the IESG to treat the technical comments in this appeal
as Last Call comments.

P1. Adequacy of WG Process
Mr. Anderson argues that the WG Chair was incorrect to declare consensus.
The IAB has reviewed the mailing list traffic during the Last Call and is
satisfied that there was indeed rough consensus to advance the -02 version
of the draft to the IESG. Indeed, Mr. Anderson appears to have conceded
this point [1]:

Rather, my interest now is to record that there were objections
raised at the time. I’m willing to acknowledge that these
objections were a minority view. I appreciate your attention
to the accuracy of the record.

Mr. Anderson also raises the issue of the WGLC having been made on the
-02 draft, whereas the -04 draft is being advanced. The changes between
these document versions are minimal in the judgement of the IAB and are
with the range of normal changes made without issuing subsequent WGLCs.

Finally, on July 1, 2006, Mr. Huston issued a WGLC [2] on the -04 draft
and called consensus on July 26, 2006 [3]. This WGLC occurred subsequent
to Mr. Anderson’s appeal to the IESG but prior to the issuance of the IETF
Last Call. In this WGLC, responses were overwhelmingly in favor of
advancement, with only Mr. Anderson objecting. Mr. Anderson’s objections
to the decision to IETF Last Call the -03 version of the document are
therefore moot.

P2. Scientific Fraud
Mr. Anderson claims that this document is based on scientific fraud in that
it relies on assertions that stateful DNS worked, made by Daniel Karrenberg
without showing supporting measurements. The IAB’s review of Anderson’s
appeal uncovers only a presentation [4] by Mr. Karrenberg in which he
presents some measurements about BGP stability and says (on slide 72):

Plea

This does not mean that anycast for DNS root service is unstable
or broken.

Please do not spread this false rumor!

However the data and its possible interpretations are judged, Mr. Anderson
does not show that the WG relied upon Mr. Karrenberg’s claims in order to
advance the document. Indeed, draft-ietf-grow-anycast-04 contains substantial
material about situations in which use of Anycast is problematic (see
Section 4), so it is clear that the WG considered this issue.

P3. Conflicts of Interest
Mr. Anderson argues that IESG members Brian Carpenter and David Kessens
and WG Chair Mr. Huston had conflicts of interest and should have recused
themselves from this decision. The IAB has reviewed the record and found
no evidence of relevant conflict of interest.

CONCLUSION
The technical issue raised by Mr. Anderson is not appealable at this time.
The IAB has reviewed the procedural issues raised by Mr. Anderson and on
the basis of the available record, the IAB concludes that the WG Chair
acted correctly in calling WG consensus and the IESG acted correctly in
issuing the IETF Last Call. We therefore conclude that the appeal should
be denied and the original IESG decision upheld.

Note: IAB members Kurt Lindqvist and Olaf Kolkman recused themselves from
this decision.

Leslie,
for the IAB.

[0] http://darkwing.uoregon.edu/~llynch/grow/msg00580.html
[1] http://darkwing.uoregon.edu/~llynch/grow/msg00462.html
[2] http://darkwing.uoregon.edu/~llynch/grow/msg00553.html
[3] http://darkwing.uoregon.edu/~llynch/grow/msg00580.html
[4] http://www.ripe.net/ripe/meetings/ripe-50/presentations/ripe50-plenary-tue-anycast.pdf