Home » CurrentHeader, Internet, Wrong On The Internet

Digital Society calls on Free Press to stop misrepresenting the IETF

By 9 September 2010 8 Comments

Free Press is publicly demanding that AT&T retract its letter to the FCC which asserted that paid prioritization was contemplated by the IETF standard.  Free Press claims that the IETF disputes AT&T’s claim by citing comments from IETF chairman Russ Housley.  Housley claimed that AT&T’s assertions about the IETF standards were misleading and that the IETF doesn’t imply any support for payment based prioritization.

The problem with Free Press’ attack on AT&T is that Russ Housley made it clear that he was not speaking on behalf of the IETF and that he wasn’t representing IETF consensus.  Not only that, Russ Housley’s comments about the IETF Differentiated Services (DiffServ) standard clearly conflict with what is written in the IETF standards.

Housley insists he wasn’t speaking for the IETF

Shortly after the story broke last Thursday on CNET that the IETF called AT&T’s assertions about IETF standards supporting paid prioritization “misleading”, Russ Housley sent an email to the IETF to say that he wasn’t trying to represent the IETF and that the reporter failed to accurately attribute his comments.  Housley wrote:

“I want the whole community to be aware of the comments that I made to the press yesterday. Clearly, these comments do not represent IETF consensus in any way. They are my opinion, and the reporter was told to express them as my opinion.”

Who was masquerading as the IETF?

Digital Society approached CNET reporter Declan McCullagh about Housley’s claim that he had been asked to express Housley’s comments as personal opinion.  We asked him if and why he disregarded this request and portrayed Housley’s comments as being representative of the IETF.  McCullagh replied with the following comment:

“As a reporter, I’m a neutral party in this dispute. (On a practical level, if I were not, then who would talk to me?)  My initial contact with Russ on this topic came after an outside PR firm contacted me. They set up our subsequent phone conversation. The email, below, might help clarify matters.”

McCullagh included the following email from the PR firm FD.

Subject:        IETF Chair speaks on Paid Prioritization

Date:   Thu, 2 Sep 2010 14:17:21 -0400
From:   Melissa Kahaly <Melissa.Kahaly@fd.com>
To:     <declan.mcCullagh@cbs.com>


Given your recent story “AT&T: Net rules must allow ‘paid prioritization”, I wanted to share with you the below comment from Russ Housley, Chair of the Internet Engineering Task Force (IETF). Please let me know if you have any interest in speaking with Mr. Housley or if you have any further questions.


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

This letter seems to indicate that McCullagh’s story accurately reflected the information that he was given.  That information was wrong but the email from FD clearly made the effort to represent Russ Housley’s views as if they came from the IETF.  But who was really trying to discredit AT&T and who hired the PR firm?

Digital Society asked Russ Housley who paid FD to reach out to the reporters and he responded that it was the Internet Society (whose front page also promotes a Russ Housley conference).  So it appears that it was actually the Internet Society and Russ Housley that wanted to discredit AT&T, but they presented Housley’s comments to the media as if it had come from the IETF through a third part PR firm.  Furthermore, the strategy worked as the story of the IETF discrediting AT&T grabbed hold in the news cycle and it enabled Free Press to go on the attack.

Russ Housley misrepresented the IETF’s position on paid prioritization

Even though Russ Housley is the current chairman of the IETF, he clearly misrepresented the IETF’s position on paid prioritization.  Housley insists that the IETF doesn’t imply any support for payment based prioritization, but this goes against what the IETF Differentiated Services (DiffServ) standard lays out in clear writing.  Declan McCullagh noted:

A July 1999 IETF specification (RFC 2638) discusses paid prioritization by saying: “It is expected that premium traffic would be allocated a small percentage of the total network capacity, but that it would be priced much higher.” Another specification (RFC 2475) published half a year earlier says that setting different priorities for packets will “accommodate heterogeneous application requirements and user expectations” and “permit differentiated pricing of Internet service.” (An RFC is a policy document, often accepted as standards, published by the IETF.)

I would also add that the abstract of RFC 2474 says:

“Differentiated services enhancements to the Internet protocol are intended to enable scalable service discrimination in the Internet

Not only did the IETF contemplate paid prioritization, it’s obvious that the term “differentiation” in DiffServ applies to payment as well as performance.  It would be silly to suggest that higher performance characteristics shouldn’t be contingent upon higher payment.  If one could acquire higher priority at no additional cost, then everyone would have higher priority which effectively means no one has higher priority.  The entire DiffServ engineering standard would be moot if didn’t accommodate the basic economics.

The DiffServ standard recommends the dropping of unauthorized priority packets

Updated 4:00 PM – This section was clarified based on Richard Bennett’s comments.

Free Press, Russ Housley, and many other Internet regulation advocates are misrepresenting the DiffServ standard.  They assert that the DiffServ standard requires network operators to blindly follow the user’s requests for prioritization without any regard for the negotiated rates.  But internet standards architect Richard Bennett, a Research Fellow at ITIF who runs the High Tech forum, pointed out to Digital Society that the DiffServ standard clearly states that unauthorized priority market packets should be discarded.  RFC 3246 which describes how DiffServ should prioritize packets between two networks states:

“To protect itself against denial of service attacks, the edge of a DS domain SHOULD strictly police all EF marked packets to a rate negotiated with the adjacent upstream domain. Packets in excess of the negotiated rate SHOULD be dropped.”

It is important to note that the IETF does not prohibit or require fees for priority packets.  The DiffServ standard only spells out that both parties must honor whatever agreement they negotiated or expect their unauthorized priority packets to be dropped.  The market place can negotiate prices and those prices could even be zero if both parties agree.  Network operators usually charge more for network connectivity with higher priority, and high priority packets are usually limited to a small percentage of the maximum bandwidth rate.  In certain instances when two network operators feel that they can offer each other equal value, they might agree not to charge each other for prioritization up to some negotiated rate.

Packet prioritization is no different than first class airline tickets or priority parcel delivery.  Even the publicly operated US Postal Service demands higher rates for priority packages.  If a customer slaps a priority label on their package but only paid for standard delivery, would we expect the USPS to delivery that package with priority service?  If a person with a coach ticket sits down in a first class seat on an airplane, the airline has the obligation to its higher paying customers to enforce seating.


Free Press is misusing the misinformed personal opinion of IETF chairman Russ Housley to attack AT&T.  Russ Housley now clearly states that his comments regarding AT&T did not represent the IETF or the IETF consensus.  It was not the IETF claiming that AT&T misrepresented the Differentiated Services (DiffServ) standard; it was the Internet Society masquerading as the IETF through a PR firm that tried to discredit AT&T.

Furthermore, Free Press and Russ Housley are misrepresenting the DiffServ Internet standard when they claim that DiffServ doesn’t support fee based differentiation.  Not only does DiffServ support fee based prioritization, it goes as far as recommending that all unauthorized priority packets be dropped.

Free Press has no right to be calling for a retraction from AT&T when they are the ones misrepresenting the DiffServ Internet standard.  Digital Society calls on Free Press to stop misusing the IETF name and to retract their misleading statements on DiffServ.


  • Richard Bennett said:

    This article misstates my remarks and the intent of RFC 3246. The statement: “Packets in excess of the negotiated rate SHOULD be dropped.” doesn’t refer to a payment rate, but to a packet rate.

    IETF does not mandate pricing (or lack of pricing) because to do so would be unlawful price-fixing. IETF standards steer well clear of any specific plan regarding payments; some of the older RFCs refered to Service Level Agreements, and even that wording has been discarded in favor of more neutral “Service Level Specifications.”

    Please do not cite me as claiming that IETF requires payment for DiffServ; they neither require it nor disallow it. The RFCs (which are not technically the products of IETF, but of the RFC Editor) are not guidelines for commercial contracts, they’re voluntary agreements about how packets look on the wire or over the air.

    Free Press has misstated the intent of the RFCs, and let’s not compound the error by misstating in the other direction.

    My comments to the FCC on DiffServ can be found here: http://www.itif.org/files/8_10_2010_ITIF_Broadband_Regulation.pdf

  • George Ou (author) said:

    @Richard Bennett,

    I never intended to say that the DiffServ standard requires payment and I’m sorry that I implied it. I have clarified my statement with due credit to you.

  • Richard Bennett said:

    Thanks, I think your update makes the situation a lot more clear. What may not be obvious about Housley’s statements is that standards engineers will often pick on a nit in a statement to criticize an argument rather than address the argument itself.

    Housley’s complaint with AT&T is not about the argument that the DiffServ architecture and standards permit carriers to charge for services, its with some obscure detail of the standards process that’s too arcane for the average citizen to appreciate. He apparently thinks AT&T says that DiffServ providers are required to charge a specific fee for the service; he doesn’t seem to be aware that AT&T was rebutting a claim by Free Press, or even what Free Press’s claim was. He regards the Free Press claim that DiffServ must be a free service as too insubstantial to mention, apparently.

    The people at IETF who actually had something to do with the DiffServ standards (which were completed long before Russ took over the chairmanship) agree that AT&T’s reading is much closer to correct than the Free Press reading, although neither is perfect in terms of standards-engineer-inside-baseball-nit-picking over terminology.

  • George Ou (author) said:

    It’s really easy to mess up the finer points and I’m sure if you speak to the people who wrote the AT&T letter regarding DiffServ, they would agree with you that their wording could have been better.

    But that is a fundamental difference here. If you tell me that I misspoke, I will acknowledge it and fix it. When you point out that Free Press or Housley misspoke to a much larger degree, they ignore you.

    “He (Housley) apparently thinks AT&T says that DiffServ providers are required to charge a specific fee for the service;”

    But AT&T did not suggest that the IETF requires fees, and Russ Housley did not accuse AT&T of saying that. AT&T said that the IETF contemplated and accommodates a fee based priority system. On the other hand, Russ Housley explicitly said that the IETF standards didn’t imply fees for prioritization and that is blatantly wrong since RFC 2638 explicitly says higher prices for premium.

    In the context of how the Internet Society set out to present Housley’s views as that of the IETF to the media, it is extremely doubtful that Housley does not know that his comments are being used to mislead the media and are being exploited by Free Press. We have pointed these issues out to Mr. Housley yet he refuses to set the record straight, and I know he reads this because he emailed me. Given how politics usually works, it’s difficult believe that these things are accidental.

  • Richard Bennett said:

    Housley’s efforts to clarify his thinking about DiffServ are ongoing. People should understand that he’s a security engineer who spends most of his time working on standards for public keys and that sort of thing; DiffServ is very far outside his expertise.

  • George Ou (author) said:

    @Richard Bennett

    His ongoing comments http://www.broadcastingcable.com/article/456906-Housely_IETF_Has_Taken_No_Position_On_AT_T_Prioritization_Assertions.php are getting weirder and weirder.

    “The problem, Housely says, is that the IETF specification at issue is not about “prioritization,” but about quality of service.”

    Yikes, there are different priority levels within Quality of Service.

  • Richard Bennett said:

    Right, prioritization is one way to provide Quality of Service. The AT&T letter actually talks about QoS anyway, and DiffServ is required to support IP Precedence, which is a system of prioritization according to Internet standards.

    Once again, we see Free Press blatantly misrepresenting the Internet’s architecture and protocols, and calling all who disagree liars and thieves.