Dr Peering -

You seemed to have stirred up quite the flame war with your Paid Peering and Net Neutrality article.  What is going on?

Felix

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

It appears that at least one person is under the mistaken impression that DrPeering went to the FCC to lobby for the regulation of peering, paid or otherwise.  This could not be further from the truth. On the contrary, DrPeering, very publicly, raised a warning flag that section 106 of the NPRM describes a rule that appears to come strikingly close to the heart of peering.

As with all of these discussions, there are two sides to the coin. Let’s clarify and focus on the core of the issue.

Policy friends and indeed Jon Peha (Chief Technologist, FCC) is cited to have said here that sometimes these rules are intentionally vague in order to elicit comments. So since :

1. a)the section 106 rule appears to have the intent of disallowing access networks to charge for “enhanced” access to the eyeballs, and 
   

2. b)paid peering enhances performance (lower latency to the eyeballs, more control over routing, more visibility, etc. by bypassing the transit providers route normally in the path),
   

raising the potential regulatory threat posed here seems appropriate. 

At the heart of the discussion is the question, “Does this rule apply to peering?”

The folks who say “No” point to the FCC diagram in section 106 (see below), shown on top of an illustration we will discuss later.

The double arrowed line between the end user and the broadband service provider appears to mean that the rule is intended to limit what an access provider can do within it’s its own network. The belief therefore is that this rule does not apply to what the access provider does at its borders where it peers.

If this is indeed the FCC’s intent, then we can all go home and sleep soundly knowing that there is no regulatory interest in peering.  (The FCC would not comment on this point when the question was asked directly. Apparently this silence is a fairness issue - if they clarify for anyone, they need to clarify for everyone, which would require a new release of the NPRM...  So if anyone claims the FCC has offered clarification, then let’s cite the source and be done.)

However, the decision to peer or not to peer at the access provider edge, in effect, makes a big difference to the prospective CDN, content-heavy ISP or content provider peer.  As shown in the diagram below, we see what, in effect, looks like two-tiered access to the eyeballs they want to reach ; enhanced (paid peering) for those who pay the metered rate for paid peering to the access provider, or the more circuitous transit path for those who do not pay the access provider to direct peering.

So if you take the bi-directional arrow in FCC diagram above to include the path from the  access network peering router all the way to the eyeballs, then paid peering is a mechanism that provides that “enhanced” access to the eyeballs. At least, that is how it looks from the outside to the content provider.

So all of this fuss is about this interpretation.... if you believe that the intent of the rule in section 106 of the NPRM is to prevent access networks from charging content companies for enhanced access to its eyeballs, one could view paid peering as precisely this -- higher performance access to its eyeballs for a metered rate executed as a paid peering product, selectively sold and priced as the access provider decides. 

If I am misinterpreting the intent, then we can all breathe a sigh of relief; peering is no where close to the regulatory scope. 

If there is any correlation between the rule and paid peering we are all in for an adventure as this will affect network relationships worldwide and may invite further regulation on a country by country basis or a network by network basis or some unforeseen future consequences.

Dr Peering

DrPeering@DrPeering.net

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

Addendum - NANOG 48 in Austin TX Feb 23, 2010 included a Net Neutrality session with excerpts from the Q&A section transcribed below.

It is significant since Jon Peha (FCC) is consistent at NANOG as in private discussion with individuals in the industry.

1. (1)When asked “Are an access network’s upstream management practices, through their transit or their peering relationships, out of scope to the NPRM rules?” the answer was that “the NPRM does not have an answer to the question”, so the issue is in other words is ambiguous.
   

at the same time...

2. (2)Dan Golding said if  you as an access provider “evaded” the spirit of the FCC rules with the network’s handling of its upstream relationships (prioritization,prepending, etc.), then the FCC would be “slapping you down”. So even if not explicitly articulated in the NPRM, under the right circumstances, an access provider’s use of paid peering for example may be brought into examination at the FCC. According to Dan anyway.
   

If the concern is that the FCC might be interested in peering and transit relationships, the NPRM is silent on the issue but Dan Golding says that even if not directly specified, the FCC will use the rules if the access network is evading the spirit of the rules (to prioritize one’s own services over those of competitors for example).  Seems like that would make peering in scope.  Too bad the FCC folks didn’t explicitly agreed with Dan Golding's statement, but alas there was no comment here from the FCC.

It seems that both side declared victory, that their interpretation was correct.  Here is the edited transcript that only includes the Q&A portion of relevance. Hopefully this will be helpful in dispassionate discussions.

Transcript

[ 10:47AM At the tail end of Questions and Answers session (Note: ‘…’ means the transcriber couldn’t hear what was said). ]

(1) Dan Golding: Um, just before we get to your question, I, I have a question that I believe is very, very important so before we run out of time which is the question of peering.  There is some discussion in the community about… Dave , would you like to pose your question about paid peering?

(2) Dave Tempkin: Sure, so the question I made reference to you earlier. You know, there’s been a lot of people who boil it down to a very, very simple question: Is paid peering really …. Am I no longer able to directly connect my network to some ther network for some undisclosed sum of money?

(3) Zachary Katz (FCC): I would say, nothing in the proposed NPRM has states that paid peering is forbidden.

(4) Dan Golding: fourth microphone

(5) William B. Norton: Well, You guys actually stole my thunder a little bit because I was going down that very specific path.

I will thank you for the clarification because I think that was the source of uh some degree of mild heated discussion.

I would ask the question a little bit more generically. Instead of saying specifically paid peering. … I actually wrote it down so um I can get it right…

Are an access network’s upstream management practices, through their transit or their peering relationships, out of scope to the NPRM rules?

(6) Dan Golding: Did you say carrier or peering?

(7) William B. Norton: Um, well, it goes to where the boundary is in that access network, and whether he sends traffic through a circuitous path or whether he goes through direct. Is that out of scope in the NPRM?

(8) Zachary Katz (FCC): I may not have the technical background to answer this, so with that caveat, I don’t believe the NPRM prohibit um..

(9) Jon Peha (FCC) – The simple answer is the NPRM itself does not, does not have an answer to that question, and that it is not explicitly….

(10) William B. Norton: and that is the answer that gets us in trouble Jon

(11) Jon Peha (FCC): No, no. This is.. If there are people in this room who think they have concerns  one way or the other, they are welcome to submit them, in fact are encouraged to help us understand the issue, but there is also to be clear, I’m not saying this is, you know,… to repeat Zach’s point, there is nothing in there that discusses these peering arrangements.  There are clear principles about access or not, uh…, and if you think that there are peering issues, then you ought to submit them in form(?).

(12) William B. Norton: No, I was looking for the answer that I thought I heard which was “No, that ‘s out of scope”, or “Yes, peering and peering practices may be a way that one would implement some of these practices that you are trying to make sure don’t happen” and that’s where the heat tends to come in.

(13) Dave Tempkin: There’s a diagram in the NPRM actually that is pretty compelling, I forget the page that it’s on, that basically shows…

(14) Dan Golding: 42, page 42

(15) Dave Tempkin that basically shows what’s in scope, so It is pretty clear that it is intended to be in the last mile, but whether or not the language needs to be cleaned up, um, you know, that remains to be seen. From the picture, that’s what it appears..

(16) Dan Golding: From my experience with the FCC that being said, ,if If folks are trying to do illicit filtering or  prioritization on peering routers , or they they try to to do things … they say “Well we aren’t doing things in the last mile but we’ll makes sure our other peoples voice traffic is degraded by getting on the peering router with routes essentially being pref’d up,  even if the NPRM right now doesn’t fit perfectly , my view of the FCC is that they are particularly humorless no offense,  when it comes to people trying to evade the spirit of the regulation. You know FCC regulations are a little bit different from most conventional law in that, most laws, if you meet the letter, the spirit matters less. With FCC regulations if you try and get around the spirit they will end up slapping you down.

(17) Richard Steenbergen: Hi Richard Steenbergen nLayer. So my question is somewhat related to what Bill was talking about. There is a liong standing practice for certain eyeball networks where they intentionally congest their transit, their peers, run their network hot. And then they say “well, if you want good access to us then you can just buy directly from us.” And this doesn’t involve any QoS because everything that you don’t pay them for is degraded . Is this something that you guys see is an issue. Because essentially what these guys are doing is using their captive customers as leverage to force other people to pay them.

(18) Dan Golding: <Dan re-asks the question…> Is that something that would be covered under the scope of network neutrality?

(19) Zachary Katz (FCC): So, it’s a real interesting point. It’s not one that I personally I’m familiar with but one I would like to learn more about before we provide guidance one way or the other.