On Fri, 9 Jun 2006, Sam Hartman wrote:

> >>>>> "Dean" == Dean Anderson <dean@av8.com> writes:
>
> Dean, here I'm speaking as an individual and I suspect some IESG
> members may agree with me and others may disagree.
>
>
>
> Dean> I misunderstood you. I thought that the "almost certain to
> Dean> appeal" meant an appeal should be accepted now.
>
> So, when there are two paths to resolving an issue--one confrontational and
> one reasonably cordial we try and choose the cordial path.

This has nothing whatsoever to do with cordial vs confrontational. This has to
do with following process. I have filed an Appeal.

> Last call can be used to resolve almost any issue including the issue of
> whether a document was incorrectly submitted to last call. In other words
> last call is for process failures as well as technical problems.

I suppose it "can be" used that way. But I have chosen to file an Appeal.

> An appeal takes a long time to deal with and is an incredibly formal and
> legalistic process.

Obviously, not. I cite my last Appeal to the IESG (3/8/2006) as an example,
where a more formal process ought to have been conducted, and indeed the formal
process was explicity specified by RFC, but not so much as a formal vote was
held. That Appeal was handled in 7 working days.

But I think these charges are serious, and do in fact warrant a formal process.

> The IESG does not consider appeal comments any more seriously than serious
> last call comments, so asking us to treat this as an appeal does not add more
> weight to your arguments it simply takes more time for us.

I fail to see why that matters: This is a serious issue, and your argument now
is that it is too much work to do the right thing?

> I agree with you that documents based on false scientific claims should not be
> published until the correctness of the document is verified using some
> mechanism that does not depend on these false claims. If the document is
> incorrect, it should be fixed. Please help all of us do that.

First, there is no way to fix this research, other than to withdraw it and
apologize for its publication. It is impossible to invent a "fix" for this
research, any more than the one could "invent" new stem cell lines to "fix" the
South Korean research that claimed to have done so. A lie written on paper
can't be undone. It can't be fixed. It is part of the historical record of
frauds.

Second, there is no way to fix this document. The assertion that you can do
Stateful Anycast TCP is absurd techincally, from first principles, as I and
others have been explaining for some time, and as noted in RFC's 1546 and 1812
etc. The Verisign collected data shows experimentally that this BGP scheme does
not work, and the previous RFC's on the topic indicate that stateful Anyast is
impossible due to axioms of packet delivery articulated in RFC1546: One cannot
assume that two packets will be delivered to the same Anycast host, therefore
Anycast is a stateless service. RFC1812 also permits fast load balancing
schemes.

Recently, we learned the "research" refuting all that never even tested
stateful Anycast.

The operations community has been sold a bill of goods that are only in the
financial benefit of the root server operators, but doesn't work, and is
impossible to make work.

TCP cannot work with two different hosts having the same IP address.
--------------------------------------------------------------------

And that is the crux of the fraud. ISC and RIPE have given assurances that, in
spite of RFC1546 and RFC1812, some BGP configuration can make this work. But in
fact, it can't. In fact, they didn't test it. In fact, they have no basis for
their assurances. In fact, they lied. In fact, they tried to suppress the
critics of their scheme. In fact, they subverted IETF and IESG processes.

http://www.av8.net/IETF-watch/DNSRootAnycast/History.html

> Dean> But I think there are enough process and other issues with
> Dean> the document to say that it should have never gotten to Last
> Dean> Call.
>
> That may be the case.
>
> The IESG can conclude that a document incorrectly made it to last
> call. Normally that happens because we realize a mistake was made.
> It could happen because we realize that an IESG member or WG chair
> intentionally subverted the process; if that happened it would be a
> very serious realization.

This sounds very much a like a process failure which may be appealed. I have
filed such an appeal.

> I think we all like to try and avoid making such determinations;

I fail to see why I should be concerned about these things. As I see it, your
duty as an IESG member, first and foremost, is to serve the public trust. If an
IESG member has subverted the process, that is a significant breach of the
public trust. Your duty seems clear: to to protect the integrity of the ISOC,
IESG, and IETF.

> we would rather fix documents than assign blame to individuals.

I think process machinations to protect certain individuals is inappropriate.

> I understand that you have made such a claim in this instance. I take no
> position on that claim because I have not researched it enough to have an
> informed opinion.

Are you saying you did not read the Appeal submitted to the IESG 3/8/2006? The
IESG responded to this appeal on 3/20/2006 with a decision.

> Dean> Particularly, since there hasn't been a WGLC on this
> Dean> draft, there is a process failure that indicates that it
> Dean> shouldn't be in Last Call.
>
> You know, I used to think that working group drafts needed WG last
> calls. I wish they did. However it turns out if you look at section
> 7.4 of RFC 2418, WG last calls are not required.

Yes, rough consensus is required, and no further messages after Jan 27 were
issued on the subject, before the IESG notice. Rather than consensus, there
seems to be lack of interest. Indeed, there have been only 6 people (besides
myself and the WG chairs) posting on the subject during its _entire_ discussion.

> Section 7.5 of RFC 2418 does require that the WG determine there is rough
> consensus to publish a draft. Without a WG last call it is usually difficult
> to make this determination. If a draft were forwarded to the IESG without a
> wg last call, then it would be easier to claim that there were not consensus.

Indeed, it does seem difficult, and there seems to be no evidence whatsoever
that there was a consensus. (I don't have to prove there wasn't consensus, the
WG chair has to _honestly_ demonstrate that there is a consensus). Its unclear
how that honest belief is possible without even a single message. And I also
note here that the WG chair works for APNIC, which was the first ISC customer of
this "Anycast" service.

And in any case, it requires an approval from the A.D. to proceed. That A.D. was
David Kessens, and Kessens was aware of the fraud.

We have 2 irregularities that should be sufficient:
1) No further email on the subject (e.g. "Are we Good with the
draft-3?")
2) Kessens knew of the problems.

But, actually, we have 3 _more_ irregularities.
3) Kessens again tried to suppress this discussion.
4) Carpenter tried to suppress my Appeal.
5) Huston works for APNIC, which is ISC customer of this service.

These five irregularities should be more than sufficient.

> Note, however that there does seem to be a WG last call on the 02 version of
> the draft ; I point to a message below where you claim you made comments
> against the draft during that last call.
> http://darkwing.uoregon.edu/~llynch/grow/msg00460.html

This last call failed, and resulted in Draft 3.

> Perhaps you are concerned that there was no last call on 03 (I don't
> know if such a last call happened.) You even admitted that the WG
> decision was made in
> http://darkwing.uoregon.edu/~llynch/grow/msg00462.html .

You are taking my note out of context. I was responding to what Huston had
previously written: http://www.uoregon.edu/~llynch/grow/msg00461.html
December 1, 2005 Geoff Huston writes:
==================================================
I have seen no further working group postings on this draft during this
last call period, and no observed working group consensus to create another
revision of this draft.
==================================================

So, I was under the impression that Huston had made a decision and was not
interested in any further discussion, even though I had documented 6 unaddressed
problems.

But it is clear now that in fact no decision had been made in December, and I
was wrong to think it had been. That is not an "admission of decision".

On January 27, 2006, another draft was issued, according to the actual group
consensus which was to create another draft.

Then we found that the assurances given by ISC were false and based on false
research.

> What has happened since then? (This is a serious question on my part) Did it
> have something to do with the research you claimed was discredited?

There is a long history of controversy:
http://www.av8.net/IETF-watch/DNSRootAnycast/History.html

To pick it up recently:

Only two messages to GROW, among other activities:
January 17, 2006 (RIPE provides source with caveats to Anderson)
January 25, 2006 (RIPE concludes agreement not to redistribute)

January 27, 2006 Draft 3 announced.
http://www.uoregon.edu/~llynch/grow/msg00517.html

February 16, 2006 (Anderson makes queries about Stateful testing in DNSMON)
March 3, 2006 (Anderson repeats queries re Stateful testing in DNSMON)
March 8, 2006 (Anderson notifies IESG of research problems with Anycast)

June 2, 2006 IESG announcment
http://www.uoregon.edu/~llynch/grow/msg00532.html

> Dean> The assertion by certain IESG members that one should
> Dean> overlook scientific fraud discredits their integrity, and to
> Dean> some extent discredits the integrity of the IESG by
> Dean> association. But when the IESG accepts such assertions with
> Dean> full knowledge, then the integrity of the IESG is certainly
> Dean> discredited.
>
> I agree that it would be inappropriate for the IESG to publish a
> document without carefully considering allegations of fraud. We have
> not done so. We will consider your allegations. You can help us do
> so by engaging in a constructive dialogue and helping people like me
> who don't understand the situation do so.

Why shouldn't that be done as an Appeal?

> We may have questions; if so, we'd appreciate your help answering
> them.

Absolutely. So far, I've only gotten resistance to considering this issue.

> We should no more blindly accept assertions of fraud than we should ignore
> these assertions. That means we cannot simply take your word that there is
> fraud. Instead, we have to go look at it and make up our opinions. We must do
> that before we approve the document.

Well, the IESG has known of these allegations since March 8th, 2006. It has had
plenty of time to "check things out", and in fact, claims to have done so by
deciding on the 3/8/2006 Appeal.

> But it is reasonable for us to conduct a last call and to collect evidence for
> both sides of the issue during the last call.

I don't think it is reasonable to ignore my process filing. I have filed an
Appeal, and the IESG has a duty to its hearing.

> Please help us do our job by helping us understand your claims and respecting
> that we'll ask those who disagree with you to state their case too.

They are not prevented from stating their side.

> Dean> I think it also incumbent on the IESG not to accept Last
> Dean> Call comments on a document that it already knows is based
> Dean> on a scientific fraud.
> Here we disagree. As I said I think it critical that the IESG not
> approve a document without considering claims of fraud. However it is
> never unreasonable under our process for us to accept comments.

It is unreasonable to move forward a process unreasonably. This document should
not be in "Last-Call", and I have filed an Appeal of that process move.

> Dean> The IESG has known this for at least
> Dean> 3 months, which raises some questions:
>
> When did the IESG learn of the fraud? I must have missed that message
> as the first I heard of this issue was when you filed your appeal.
> Again, I'm being serious not sarcastic. I get a lot of mail and I
> might have missed a message on this issue.

http://www.ietf.org/IESG/APPEALS/Anderson-appeal-03-08-2006.htm

> Now it may well be that David and possibly other IESG members knew about your
> claims of fraud. However an IESG member knowing something is not the same as
> the IESG knowing something. You may believe (and even be right) that IESG
> members have an obligation to tell the rest of the IESG about claims of fraud
> and that by not doing so the IESG is culpable for the actions of its members.

"Willful ignorance" and "negligent ignorance" are not a defense of any sort.
(ask Ken Lay and Jeff Skilling) We can assert that IESG knows because it was
officially notified, and the contents of the notice are on the IESG website.

We can assert that the IESG members also know the contents of the documents they
have offered opinions or votes on. So everyone who "voted" according to Brian
Carpenter, knows. That includes both Carpenter and Kessens.

Indeed, pretty much the entire IESG knew, and considered the charges in the
3/8/2006 Appeal, and the IESG issued a decision on 3/20/2006.

The Appeal mentions Kessens participation by name.

> Thanks for your cooperation,
>
> --Sam
>
>

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