Inline

On Wed, 28 Jun 2006, Joe Abley wrote:

> Dean,
>
> These points have all been covered before. Once more, then, for the
> road.
>
> On 28-Jun-2006, at 16:57, Dean Anderson wrote:
>
> > On Wed, 28 Jun 2006, Sam Hartman wrote:
> >
> >> Dean> The -02 draft did not complete WGLC. A new draft was
> >> Dean> issued.
> >>
> >> It's often the case that a WG will decide that the result of a WG
> >> LC is that minor changes are required and that a new draft will be
> >> issued but that no new LC is required.
> >
> > I agree this can happen. This was the case with the NSID draft,
> > which also contained errors resulting from confusion and moving too
> > fast. But that wasn't the case for the draft-ietf-grow-anycast-02
> > document.
>
> In fact, it was. Your statement is incorrect. There are more details
> below.
>
> > In that case, the Last Call issued in November by Geoff Huston also
> > indicated that there were no comments requiring a new draft. This
> > was wrong.
>
> The last time you voiced this opinion, the grow chair replied:
>
> http://darkwing.uoregon.edu/~llynch/grow/msg00461.html
>
> After the WGLC had been concluded, Geoff forwarded the document to
> David Kessens as AD requesting that the document be published. David
> Kessens contacted the authors of the drafts with a list of
> typographic, spelling and wordsmithing complaints. We fixed those and
> resubmitted it as -03, as requested by David.

On the issue of events:

The datatracker and the GROW WG list dispute Abley's series of events:

2005-11-21 Huston says no comments and no new draft need.
http://darkwing.uoregon.edu/~llynch/grow/msg00449.html

2005-11-22 Draft Added by Dinara Suleymanova in state Publication
Requested

2005-11-30 Anderson disputs Huston on no comments claim. Notes that
there were posting against this draft.
http://darkwing.uoregon.edu/~llynch/grow/msg00460.html

2005-11-30 Huston acknowledges comments, again asserts consensus and no
new draft. http://darkwing.uoregon.edu/~llynch/grow/msg00461.html

2005-12-02 Anderson asserts Huston misinterpreted, the discussion was
not over, and summarizes the 6 complaints against the draft.
Apparently, this message has some effect. Though, it wasn't very clear
at the time, even to me (Anderson), that my message would have any
effect. But it did. The next event is that the state was changed from
"Publication Requested" to "Revised ID Needed"
http://darkwing.uoregon.edu/~llynch/grow/msg00462.html

2005-12-21 State Changes to AD is watching::Revised ID Needed from
Publication Requested by David Kessens

2006-01-27 03 [system] New version available

This new -03 draft has significant changes, which require a new
consensus.

On the issue of "minor changes":

This is a major change:

- discrete locations. The service provided by each node is consistent
- regardless of the particular node chosen by the routing system to
- handle a particular request.
+ discrete locations. The service provided by each node is generally
+ consistent regardless of the particular node chosen by the routing
+ system to handle a particular request (although some services may
+ benefit from deliberate differences in the behaviours of individual
+ nodes, in order to facilitate locality-specific behaviour; see
+ Section 4.6).

The change from "is consistent" to "is generally consistent" is a
significant difference. To be a safe stateful operation, "is consistent"
is required behavior. You assured people that it is safe, and then you
"bait and switched" the text to weaken the official statements in spite
of your stronger assurances. This is a nefarious change.

This is a major change:

- Some anycast deployments have very predictable routing systems,which
- can remain stable for long periods of time (e.g. anycast within an
- well-managed and topologically-simple IGP, where node selection
+ Services may be anycast within very predictable routing systems,
+ which can remain stable for long periods of time (e.g. anycast within
+ a well-managed and topologically-simple IGP, where node selection

While this may at first seem to be simple change, it de-emphasizes the
fact that some routing systems complying with RFC 1812 don't have
predictable patterns. That is, they may use fast load switching such as
PPLB. This is a nefarious change.

(I stopped looking here, having made my point)

> A unified diff between -02 and -03 is included below.

Thanks. It shows many significant changes.

> > A new draft was then issued.
>
> -03 was forwarded to the IESG for publication as BCP, and was
> subsequently subject to Secdir and Gen-ART review. Those reviews
> resulted in changes which were greater in extent than the simple typo/
> spelling/wordsmithing corrections from -02 to -03. Those changes were
> incorporated into what will one day hopefully be called -04, the
> current text of which has been sent to the grow list as -pre04.
>
> The IESG issued a last call on -03 and as far as I know they received
> no comments on the draft apart from yours.
>
> > The November last-call failed to achieve consensus on the -02
> > document. Therefore, the draft-ietf-grow-anycast-02 document did not
> > "complete the last call", as Abley asserts.
>
> That's not correct. See, for example:
>
> http://darkwing.uoregon.edu/~llynch/grow/msg00451.html

This 2005-11-22 call by Huston was disputed at the time. In the message
you quote, Huston incorrectly claims there were no comments and no new
draft needed. In fact, as I demonstrated at the time, there were
actually a number of comments. Huston's assertion of consensus was
similarly incorrect. Obviously, Huston didn't consider these comments
properly on 11/21/05, but did consider them by 12/21/05.

A failed last call is not "complete".

> >> Another common reason for a post-wg-LC draft is comments received
> >> from the AD between the time when the AD accepts the draft and an
> >> IETF last call is issued.
> >
> > I understand that is possible, too. But it is not the case for the
> > draft-ietf-grow-anycast-02 document.
>
> Again, that's factually incorrect.

The datatracker doesn't support your version. Also, such post-LC
changes are minor, usually along the lines of spelling, grammer,
clarity. No substantantial changes are allowed. Version 03 has a lot
of changes from version 02.

From 2006-01-27, when the new version was available, there passed 5
months, during which there was no comment on the draft. The draft needs
a new consensus, especially given the extraordinary new facts which were
learned by the IESG and yourself during this time.

http://darkwing.uoregon.edu/~llynch/grow/msg00451.html> As I mentioned
in my recent message to the grow list, once we are able
> to submit -04 we will ask the grow wg chair to carry out a new grow
> wglc specifically on the changes made to the document between -02 and
> -04. If the chair agrees to this, then there will have been no
> revisions (minor, typographical, or otherwise) which have not been
> subject to a working group last-call.

That seems premature, and inappropriate. But more to the point: How do
you know anything about the IESG decision on this document? And why do
you think that the IESG won't publish the document as is? There seems to
be something inconsistent in your actions, here.

Given knowledge of the scientific fraud, I think there is no chance you
document will be published, because certainly no one will just accept
your assurances that stateful anycast works and is stable. And only a
brief glance at
http://www.av8.net/IETF-watch/DNSRootAnycast/History.html demonstrates
that this was never an uncontroversial issue, as you and your associates
have also asserted. And we have the data that shows that stateful
anycast is unstable. How do you expect to get approval for an unstable
activity associated with a scientific fraud?

> To the best of my knowledge, there is not a single person who has
> read this draft, apart from you, who objects to its publication. The
> following may provide useful background:

Actually, of the 7 people total not including authors and chairs who
posted on the subject, 2 disagreed with the content. The other 5 mostly
talked about BGP terminology and whether PPLB existed [Nanog posters had
said that it wasn't possible on GSRs--and while I also believed these
comments at the time, those claims too, was found to be untrue: GSR E2
cards do PPLB on all interfaces or none. You can't do PPLB only on
"parallel links" (All unicast links are parallel, but what is meant is
single-hop parallel links.)] Verisign's data also contradicts this
claim---at the time Karrenberg had authoritatively disputed Verisign's
data, but we now know that his claims were false. That's hardly a broad
consensus. And all of these people were previously given unqualified
assurances by ISC and RIPE that stateful anycast was safe and stable,
and widely used. None of those assurances were true. Given that these
assurances are now known to be fabricated and false, I think we can
safely assume that few if any of these 5 people would now agree to
publish your document. And I think even if they did (there are after
all more conspirators at ISC and RIPE than there are GROW posters on the
subject), I suspect that the IESG would have questions.

To the best of my knowledge, there is not a single honest person or
honest organization that would _knowingly_ publish a document that
furthers a scientific fraud such as stateful anycast. This seems to be
a distinguishing feature of having or not having "honesty" and
"integrity".

In fact, your camp is singularly unique in arguing that it is somehow
too late to reject a document based on a scientific fraud. What Kessens
did in June 2006 was entirely inappropriate, and an abuse of process.
Kessens fabricated statements in the datatracker comments for this
document. Kessens unquestionably knew of the fraud by June. Indeed,
there is evidence that Kessens was an active participant in the fraud
from at least Septemer, 2005, possibly earlier. Even if there _was_ a
completed last call in November 2005, and no changes to the document
between -02 and -03, by June 2006, it was inappropriate to try to
publish this document. Your argument about "completed last call" is
entirely moot except for the question of its historical accuracy.

Well, actually, there is some utility to your inaccurate claim. By
making this inaccurate assertion, and not telling people of the fraud
and the problems with the document, you assert a consensus where there
is none. The message is "We just missed the deadline. Everyone already
agrees so you don't need to look at this, just few finishing touches".
That is consensus building by deception. Honesty, Integrity, Full
disclosure requires something else, entirely.

Your message to the WG list
http://darkwing.uoregon.edu/~llynch/grow/msg00547.html is deceptive in
light of what's now known about the document.

> http://dictionary.reference.com/browse/consensus

Consensus through deception is a well known political practice. It is
not so common in scientific and engineering circles. It succeeds in
politics because an election is conducted before all the facts can be
actually be known, and wielding power of elected office is its own
reward, even when obtained by deception. In contrast, Science and
Engineering, and particularly Standardization, proceeds at a slower
place and depends heavilly on the truth of prior assertions.

--
Av8 Internet Prepared to pay a premium for better service?
www.av8.net faster, more reliable, better service
617 344 9000