Re: My comments to the press about RFC 2474

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



I'm making a very simple request, Brian: I want a new press release to go out that corrects the one that most assuredly did go out last week.

If you want to analyze this situation as you would an RFC, think of it this way: some RFCs have errata. We don't rely on word of mouth to correct them, we do in writing and on the record. Here are the two errors that concern me:

Error number 1: The press release didn't even include Russ' name, only his title: "IETF Chair speaks on Paid Prioritization." That clearly identifies the statement as coming from IETF.

Error number 2: The press release was at best one-sided: it falsely accused AT&T of making a claim they did not make regarding "specific payment based on prioritization as a separate service." AT&T did not say that IETF *requires* payment for DiffServ, they rebutted a counter-claim by another party to effect that charging *is not allowed* by RFC 2474. AT&T's actual claim was not misleading, but the unnamed IETF Chairman's characterization of it was misleading. He set up a strawman and attacked it. The truth is that AT&T claim was correct on the facts, and the other claims were wrong on the facts. RFC neither requires nor precludes charging, and in fact says: "A packet initially marked for the Default behavior MAY be re-marked with another codepoint as it passes a boundary into a DS domain so that it will be forwarded using a different PHB within that domain, possibly subject to some negotiated agreement between the peering domains."

It doesn't seem to me that an awful lot of hair-splitting is necessary to determine the correct course of action. A pair of errors were made in a press release that went out under our collective names. They need to be corrected with a new, truthful press release which is public and on the record.

RB


 
On 9/7/2010 7:03 PM, Brian E Carpenter wrote:
On 2010-09-08 11:26, Richard Bennett wrote:
  I think you should have shared the message from our public relations agency 
that started this incident, Russ. Here's what it said:
As Marshall indicated, this seems to have no public existence
outside of the present thread. However, let's assume it has gone
through some "dark side" media channels...

------------------
IETF Chair speaks on Paid Prioritization - Thursday, September 2, 2010

"I note the recent discussion in the U.S. media in connection with 'paid
prioritization' of Internet traffic and the claim that RFC 2474
'expressly contemplating paid prioritization.'  This characterization of
the IETF standard and the use of the term 'paid prioritization' by AT&T
is misleading.  The IETF's prioritization technologies allow users to
indicate how they would like their service providers to handle their
Internet traffic. The IETF does not imply any specific payment based on
prioritization as a separate service."

Melissa Kahaly
Assistant Vice President
  <http://www.fd.com/>
88 Pine Street, 32nd Floor
New York, NY, 10005
T +1 (212) 850-5709
F +1 (212) 850-5790
M +1 (732) 245-8491
www.fd.com <http://www.fd.com/>

A member of FTI Consulting Inc.
-----------------

This clearly isn't Russ Housley speaking as an individual, this is the IETF 
Chair making an official statement.
The word "official" is really overloaded. Nothing the IETF does
is official; all our standards are voluntary standards; they
all fall under a "DISCLAIMS ALL WARRANTIES" boilerplate.

So, we can discuss whether Russ was quoted in his capacity as
IETF Chair; and if so, I have no problems with that, because...

The statement is misleading as RFC 2474 neither *implies any specific payment* 
nor *denies any specific payment*. 
I really do not see any respect in which the quote from Russ is
in any way misleading. Your statement about RFC 2474 is true.
So is Russ' statement. They are simultaneously true. Russ is
simply stating what I said the other day: the IETF says nothing
about business practices or prices; RFC 2474 is an example of
saying nothing.

RFC 2475, RFC 2638, and RFC 3006 are plenty 
clear on the relationship of technical standards to commercial arrangements.
For 2475, see my comment below. 2638 is a "for the record"
publication of some pre-IETF work; it is not an IETF document.
3006 is irrelevant - it relates to intserv, not diffserv.

And yes, the Architecture RFCs are classified as "Informational" but that 
doesn't stop the Proposed Standards from referencing their "requirements" as RFC 
3246 does:

"In addition, traffic conditioning at the ingress to a DS-domain MUST ensure 
that only packets having DSCPs that correspond to an EF PHB when they enter the 
DS-domain are marked with a DSCP that corresponds to EF inside the DS-domain. 
*Such behavior is as required by the Differentiated Services architecture* [4 
<http://tools.ietf.org/html/rfc3246#ref-4>]. It protects against 
denial-of-service and theft-of-service attacks which exploit DSCPs that are not 
identified in any Traffic Conditioning Specification provisioned at an ingress 
interface, but which map to EF inside the DS-domain."
I really don't understand why you are picking this particular
nit, but you are picking it all wrong. Unfortunately this RFC
predates the practice of separating normative and informative
references, but the normative statement in the above is the
sentence containing MUST. You do not need to read RFC 2475
to understand the normative statement. Therefore, RFC 2475 is
not a normative reference. The sentence starting "Such behavior"
is explicative, not normative.

Not that it matters, anyway. Nowehere in the diffserv documents
is there any substantive discussion of pricing. The most I
can find is this single sentence in 2475:

"Service differentiation is desired to accommodate
heterogeneous application requirements and user expectations,
and to permit differentiated pricing of Internet service."

Full disclosure: I was co-chair of the diffserv WG, so
I know both the intent and the content of the documents
quite well. And as co-chair, I whacked down any attempt
to discuss pricing whenever it came up in discussion.
Excuse me shouting, but is it hard to understand that
WE DON'T TALK ABOUT PRICING IN THE IETF?
http://en.wikipedia.org/wiki/Sherman_Act

[Footnote 4] Black, D., Blake, S., Carlson, M., Davies, E., Wang, Z. and W. 
Weiss, "An Architecture for Differentiated Services", RFC 2475 
<http://tools.ietf.org/html/rfc2475>, December 1998.

I don't have any desire to limit Russ Housley's free speech rights, but it's 
clear from all the evidence that he approached the press as the Chairman of IETF 
with a statement to make about the argument between AT&T and Free Press, and 
it's the statement in the official capacity that bothers me. I wouldn't take up 
the IETF's time with a personal disagreement between Russ' interpretation of 
DiffServ and anyone else's, but this issue is clearly far beyond that.
As I see it, Russ attempted to correct misrepresentations about
IETF documents that had appeared in the press. I applaud him
for doing so.

Finally, the term "paid-prioritization" wasn't coined by AT&T, it comes from the 
statement by Free Press that AT&T was criticizing. In Free Press' usage it means 
any departure from FIFO behavior for a fee.
Ack
   Brian

RB

On 9/7/2010 3:52 PM, Russ Housley wrote:
Richard:

Russ said to the press that he considers AT&T's belief that the RFCs
envisioned payment for premium services implemented over DiffServ or
MPLS to be "invalid."
This is not what I said.  I said 'misleading.'

The letter from AT&T jumbles some things together.  AT&T makes many
correct points, but in my opinion, a reader will get a distorted
impression from the parts of the letter where things get jumbled.

Adding to this situation, it is clear to me that the term "paid
prioritization" does not have the same meaning to all readers.  If you
read the AT&T letter with one definition in your head, then you get one
overall message, and if you read the letter with the other in your head,
then you get a different overall message.  I tried to make this point.

This was captured pretty clearly in the article by Eliza Krigman:
| The feud boiled down to what it means to have "paid
| prioritization," ...

As I said on Friday, I made the point that DiffServ can be used to make
sure that traffic associated with applications that require timely
delivery, like voice and video, to give preference over traffic
associated with applications without those demands, like email.

Unfortunately, it is not simple, and I said so.  I used an example in my
discussion with Declan McCullagh.  I think that Declan captured this
point in his article, except that he said 'high priority', when I
actually said 'requiring timely delivery':
| The disagreement arises from what happens if Video Site No. 1 and
| Video Site No. 2 both mark their streams as high priority. "If two
| sources of video are marking their stuff the same, then that's where
| the ugliness of this debate begins," Housley says. "The RFC doesn't
| talk about that...If they put the same tags, they'd expect the same
| service from the same provider."

Clearly, if the two video sources have purchased different amounts of
bandwidth, then the example breaks down.  However, that is not the point
in this debate.

Russ

-- 
Richard Bennett

------------------------------------------------------------------------

_______________________________________________
Ietf mailing list
Ietf@xxxxxxxx
https://www.ietf.org/mailman/listinfo/ietf

-- 
Richard Bennett
_______________________________________________
Ietf mailing list
Ietf@xxxxxxxx
https://www.ietf.org/mailman/listinfo/ietf

[Index of Archives]     [IETF Annoucements]     [IETF]     [IP Storage]     [Yosemite News]     [Linux SCTP]     [Linux Newbies]     [Fedora Users]