Home / Blogs

Interconnection Disputes Are Network Neutrality Issues (of Netflix, Comcast, and the FCC)

A lot of people have been talking about the “interconnection” deal between Comcast and Netflix and whether that deal is related to network neutrality. (It is.) This question comes partly because the FCC’s 2010 Open Internet Order (also known as the network neutrality order) was recently struck down. So network neutrality lands back at the FCC, with a new Open Internet proceeding, at the same time Netflix starts working so poorly on Comcast that Netflix had to cut a special deal with Comcast.

Several people have argued that interconnection issues should be considered in the new Open Internet proceeding. These people include Reed Hastings, the CEO of Netflix, the Internet backbone providers Cogent and Level 3, Consumers Union, and the Internet Association (Google, Facebook, eBay, Amazon, Airbnb, LinkedIn, etc.) On the other hand, the FCC Chairman Tom Wheeler has stated”[p]eering and interconnection are not under consideration in the Open Internet proceeding,” apparently because interconnection is “not a net-neutrality issue.” My friend Harold Feld at Public Knowledge has said, “Wheeler is right, this is not a ‘network neutrality’ issue.”

Actually, I believe Chairman Wheeler and Mr. Feld are wrong, and I hope this post persuades them otherwise. Interconnection has always been a network neutrality issue.

I have two arguments on principle and one response on politics. First, on principle, interconnection has always been part of the open Internet proceedings, as evidenced by several major FCC and congressional orders. Second, ISPs can block traffic, discriminate, or impose access fees either once traffic is within their network (through “deep packet inspection”) or when the traffic is at the edge of their network (through interconnection). There is no reason to think the technical distinction should matter. Third, if the Chairman and Mr. Feld are taking politics into account, separating out interconnection from the Open Internet proceeding is an even worse idea, though it may not seem like it now.

First, the major “network neutrality” orders place interconnection front and center.

Among the FCC’s most important statements on network neutrality is a well-known Internet Policy Statement adopted in 2005. It included four principles, including that consumers should be able to access the content, applications and services, and devices of their choice. All four principles were adopted with this goal in mind, repeated four times on page 3: “To encourage broadband deployment and preserve and promote the open and interconnected nature of the public Internet.” Open and interconnected.

This Policy Statement became a two-year enforceable merger condition in both the AT&T/SBC merger and the MCI/Verizon merger.

The AT&T/BellSouth merger also included a commitment to the Internet Policy Statement and a “neutral network” offering no service that “privileges, degrades or prioritizes any packet transmitted.” In addition to accepting the Policy Statement rooted in interconnection, this network neutrality condition extended “up to and including [not excluding] the Internet Exchange Point closest to the customer’s premise, defined as the point of interconnection that is logically, temporally or physically closest to the customer’s premise where public or private Internet backbone networks freely exchange Internet packets.”

In 2008, in the FCC’s seminal Free Press-Comcast decision, concerning Comcast’s blocking of peer-to-peer applications, the FCC justified its decision in part based on ensuring interconnection. The Commission concluded that section 256, a section authorizing the FCC to set technical standards for interconnection, authorized the FCC’s action. Essentially, by blocking peer-to-peer traffic within its network through deep packet inspection, Comcast obviously undermined seamless interconnection amongst networks. The Commission concluded (on page 12): “It is therefore a reasonable exercise of the Commission’s authority ancillary to section 256 to promote the ability of Comcast customers and customers of other networks, including public telecommunications networks, to share content and applications with each other, without facing operator-erected barriers, i.e., to ‘seamlessly and transparently transmit and receive information.’”

Additionally, in 2009, Congress defined the FCC’s Internet Statement as imposing “nondiscrimination and interconnection” obligations. Congress imposed these obligations on any networks built with stimulus funding, stating “non-discrimination and network interconnection obligations that shall be contractual conditions of grants awarded under this section, including, at a minimum, adherence to the principles contained in the Commission’s broadband policy statement.” So even Congress considered network neutrality to include nondiscrimination and interconnection.

Finally, the Commission’s 2010 Open Internet Order relied partly on section 251, the section requiring interconnection of telecommunications networks. The FCC stated (on pages 69-70) that: “Section 251(a)(1) of the Act imposes a duty on all telecommunications carriers ‘to interconnect directly or indirectly with the facilities of other telecommunications carriers.’ ... To the extent that VoIP services are information services (rather than telecommunications services), any blocking or degrading of a call from a traditional telephone customer to a customer of a VoIP provider, or vice-versa, would deny the traditional telephone customer the intended benefits of telecommunications interconnection under Section 251(a)(1). ... To the extent that VoIP services are telecommunications services, a broadband provider’s interference with traffic exchanged between a provider of VoIP telecommunications services and another telecommunications carrier would interfere with interconnection between two telecommunications carriers under Section 251(a)(1).”

Said another way, ensuring interconnection has always been at the core of the FCC’s “Open Internet” mission, and explicitly so in all the FCC’s major Open Internet orders. In 2005, the FCC stated that mission in terms of consumers’ ability to access the content, applications, or devices of their choice. It first applied that mission in the Free Press-Comcast case, which involved the use of deep packet inspection to block uploads from Comcast’s own users. So the Commission simply did not have occasion in that case to address blocking and discrimination (or access fees) through interconnection with other networks. The Commission’s repeated emphasis on ensuring interconnection—including relying on it as a basis for jurisdiction twice—makes it clear that interconnection is part of the Open Internet proceedings.

Sure, the FCC’s involvement with interconnection is much older than the Open Internet orders—but that’s why the FCC knew to be concerned about interconnection from the very first of the Open Internet orders.

Second, it does not matter if carriers engage in blocking, discrimination, or access fees through deep packet inspection or through interconnection. Either way, consumers and edge providers would be affected, and innovation and free expression stifled. The Commission has generally sought to exclude agreements among backbone and transit providers that do not have termination monopolies over users, as such providers appear to operate in a competitive market. ISPs with termination access monopolies, such as Comcast, AT&T, and Verizon, were the target of the open Internet proceedings, and the DC Circuit found it reasonable for the FCC to impose rules on them because of this termination monopoly. There was no exception allowing these carriers to engage in abuse (undermining the open and interconnected Internet) if that abuse was through interconnection, not deep packet inspection or domain name blocking.

The evidence for excluding interconnection disputes is thin. The primary evidence is footnote 209 of the 2010 Open Internet Order: “We do not intend our rules to affect existing arrangements for network interconnection, including existing paid peering arrangements.” That statement, however, seems to apply only to then-existing arrangements. As a result, it seems that the Commission may have meant to address all future agreements. Moreover, this footnote attaches to a paragraph, in fact, that explicitly forbids termination access fees: “Some concerns have been expressed that broadband providers may seek to charge edge providers simply for delivering traffic to or carrying traffic from the broadband provider’s end-user customers. To the extent that a content, application, or service provider could avoid being blocked only by paying a fee, charging such a fee would not be permissible under these rules.” Since the Commission says charging the fee would be impermissible, we should consider how an ISP would assess such a fee. It would do it by threatening to block edge provider either through deep packet inspection or some other means—perhaps congestion in interconnection links or domain name blocking. It would be illogical for this footnote somehow to permit access fees for discriminatory treatment, as the Commission concluded (page 43): “it is unlikely that pay for priority would satisfy the ‘no unreasonable discrimination’ standard.” The footnote therefore is pretty weak evidence, in the face of the rest of the order and the string of FCC Open Internet decisions rooted in interconnection.

Third, the politics of separating interconnection from other network neutrality disputes is deeply flawed. Mr. Feld’s blog post suggests that the Chairman, whom he believes wants to address interconnection, is brilliantly avoiding the term “network neutrality” because he understands that this term brings political heat. The Chairman is playing chess, making a strong opening move, getting some political breathing room to analyze interconnection disputes, according to Mr. Feld. If the Chairman wants to set interconnection rules not only for Comcast (in their merger request), but also on Verizon and AT&T, he will need some political finesse.

But a great chess player also has to plan an endgame. When the end comes for interconnection, the carriers will bring overwhelming political heat. Their armies of lobbyists and lawyers will not be fooled by terms. They will know that interconnection, like network neutrality, might cost their bosses money, and one does not easily take money from Comcast, Verizon, or AT&T. When the Chairman faces this political heat, he would love to have a broad coalition and the American public on his side. During nine years of public discussion, the American public has shown over and over that it supports network neutrality. Most of the public does not know the term “interconnection.” Tech executives and lawyers who support network neutrality need to be educated on interconnection (or told it’s always been part of the same proceeding) before supporting it. And the broad coalition of companies active in D.C. might have to split resources, as Netflix, Cogent, and Level 3 perhaps focus on interconnection while other companies focus their resources on the Open Internet proceeding. I have trouble seeing a political endgame where the Chairman can rule for the public on a term that even readers of Reddit, Wired, and Ars Technica don’t understand, when the carriers will bring their usual full-frontal attack. It’s better to keep the broader coalition together and use the term that the public and tech companies understand.

Plus, network neutrality has always been about interconnection.

(Aside, if you’re wondering why anyone would avoid the term “network neutrality” in D.C., even though it is popular among the public, keep in mind some things that are popular in D.C.: telecom lobbyists, oil companies, mass surveillance, contractors that build stuff like healthcare.gov, and investment banks. Network neutrality is popular outside of D.C. but it’s sometimes hard for folks in D.C. to remember that so they talk of interconnection.)

By Marvin Ammori, Fellow at the New America Foundation, Lawyer at The Ammori Group

He advises a range of technology clients, large and small. The views expressed here are his own and should not be attributed to any other institution or client.

Visit Page

Filed Under


Separate. Related. Michael Elling  –  Apr 9, 2014 3:49 PM

2 views illustrate how interconnect and NN are separate yet related:
1) where is the MAN/WAN demarc? This is a horizontal boundary question.

2) what is the tradeoff between layer 2 transport and layer 3+ switching and storage? This is a vertical boundary question.

Frame your terms, technology and costs/prices along those boundaries.

The debate is centered around HD 1-way content which is the past and present. It should instead be focused around 4K 1-way content, HD 2-way collaboration, increased mobility and IoE/IoT; the future.

All of these, with 1-way HD content leading the charge, call for the boundaries to shift out to the edge and up in the stack in terms of distribution of intelligence to the edge. That’s how we keep costs down and universal service up, which seems to be every policymaker’s challenge these days to capture the bottom quintile of demand, be it defined by income, teledensity, or usage.  This perspective will benefit all stakeholders, even the incumbents, if viewed this way.

Comment Title:

  Notify me of follow-up comments

We encourage you to post comments and engage in discussions that advance this post through relevant opinion, anecdotes, links and data. If you see a comment that you believe is irrelevant or inappropriate, you can report it using the link at the end of each comment. Views expressed in the comments do not represent those of CircleID. For more information on our comment policy, see Codes of Conduct.

CircleID Newsletter The Weekly Wrap

More and more professionals are choosing to publish critical posts on CircleID from all corners of the Internet industry. If you find it hard to keep up daily, consider subscribing to our weekly digest. We will provide you a convenient summary report once a week sent directly to your inbox. It's a quick and easy read.

I make a point of reading CircleID. There is no getting around the utility of knowing what thoughtful people are thinking and saying about our industry.

Co-designer of the TCP/IP Protocols & the Architecture of the Internet



New TLDs

Sponsored byRadix

Domain Names

Sponsored byVerisign

Threat Intelligence

Sponsored byWhoisXML API

Brand Protection

Sponsored byCSC

IPv4 Markets

Sponsored byIPv4.Global


Sponsored byDNIB.com


Sponsored byVerisign