The Internet Model

I filed comments with the FCC in the Internet Freedom docket suggesting that the FCC invest some time in understanding how the Internet model of multi-stakeholder governance can help. The Commission needs to develop a process for investigating and resolving net neutrality complaints that is more effective than the “three bright line rules and some prayers” approach it’s taken in recent years. I provide an account of the errors the FCC made in responding to the Free Press/Public Knowledge petition in 2007 over Comcast’s treatment of P2P. This brief history explains that the response of the Internet community to the issue was much more effective than the FCC’s ham-fisted assessment.

Subsequent FCC actions – the 2010 and 2015 Open Internet Orders – simply magnified the issues that came to light in the Free Press/Public Knowledge incident as the FCC merely tried to find a lawful way to make a hash of issues it never understood.

My full comments are available for download here.

The Internet Model

The Internet Model of governance is a multi-stakeholder approach adapted to the Internet’s unique circumstances. Former NTIA Administrator Lawrence Strickling described the benefits this model has provided to citizens through its application to the Internet:[1]

Today the world’s citizens are benefitting from the growth and innovation of the Internet. The Internet has flourished because of the approach taken from its infancy to resolve technical and policy questions. Known as the multistakeholder process, it involves the full involvement of all stakeholders, consensus-based decision-making and operating in an open, transparent and accountable manner. The multistakeholder model has promoted freedom of expression, both online and off. It has ensured the Internet is a robust, open platform for innovation, investment, economic growth and the creation of wealth throughout the world, including in developing countries.

More importantly, Strickling touted the superiority of the Internet model over traditional telecom regulation:

Decentralized control over the Internet involving innovators, entrepreneurs and experts is far preferable to a top-down government approach that has political dealmakers charting the future of the Internet, especially for citizens of countries in the developing world. We need to work together to chart a course beyond Dubai that considers these matters in suitable multistakeholder venues so that discussions are well informed by the voices of all interested parties. Our shared commitment should be to ensure that our respective citizens benefit from the Internet and our USTTI bonds provide a solid foundation for us to chart a path forward together.

The Internet is a dynamic system, constantly evolving, improving, facing new challenges, enabling new forms of interaction, and extending infrastructure based on new technologies to new users and new areas. Because the facts are constantly changing, the traditional top-down regulatory model is difficult to apply to the Internet.

The top-down model presumes that the regulator makes a finding of fact and applies the pertinent regulation. While this process is straightforward in a milieu of slow change or no change, such as the telephone network space, it has always been difficult to apply to the Internet.

The FCC’s infamous 2008 Comcast Order illustrates these difficulties.[2] This matter involved a complaint about a network management practice briefly used by Comcast for the purpose of protecting those of its customers using the Vonage VoIP telephone service from call degradation.

The practice lent itself to colorful description. While it consisted of reducing the number of virtual circuits used by P2P file transfer programs BitTorrent and Vuze, the manner of the reduction involved injecting “TCP Reset” packets into the customer data stream. Hence, critics declared Comcast was “impersonating” users.[3]

Other critics claimed Comcast was deliberately degrading P2P in order to protect its MVPD business.[4] But the reality was that Comcast deployed a stopgap system to protect users of VoIP after the order it had placed for upgraded DOCSIS 3 CMTS systems was put on hold by the vendor because of a delay in the development of the DOCSIS 3 standard.[5]

By the time the FCC issued its order, Comcast had discontinued the TCP Reset practice in favor of an application-neutral approach to congestion management known as “Fair Share”. The Fair Share system was shared with the Internet Engineering Task Force and published in RFC 6057.[6]

The Fair Share system didn’t solve the entire problem that P2P created for VoIP; the problem was actually triggered by “buffer bloat”, a widespread issue with router design that was not well understood for many years.[7]

BitTorrent modified its own code to reduce the impact it had on other applications, again sharing the approach with IETF in the LEDBAT RFC.[8] IETF went on to develop an understanding of the problems inherent in controlling queue delay caused by buffer bloat.[9] Modern DOCSIS CMTS systems incorporate Active Queue Management systems that resolve the buffer bloat problem in an acceptable manner, but the issue demands additional work.[10]

The FCC’s action, an order demanding that Comcast discontinue a practice it had already discontinued, was not helpful. The Order was vacated by the courts, but the takeaway for many was to provide the agency with a stronger legal footing for what was essentially a useless, symbolic action.[11] Today, advocacy groups still raise money by complaining about the Comcast-BitTorrent matter.[12]

This is not to say that BitTorrent users didn’t have a legitimate beef with Comcast. Some uses of P2P file transfer programs are legitimate, and users should be entitled to run these programs for lawful purposes.

This matter is important because it contrasts FCC enforcement of an Internet user issue with multi-stakeholder action. The FCC’s fact-finding fell short because, in fact, no one understood the buffer bloat problem in 2007-8. The FCC’s enforcement action did not go to the heart of the problem because the agency didn’t know what the problem actually was.

To the extent that buffer bloat, inter-application side-effects, and Internet congestion are understood today, that understanding and the relevant mitigations have come from the multi-stakeholder arena rather than the realm of regulation.

Regulatory actions tend to divide the world into good guys and bad buys, white hats and black hats. But the Internet can only be the functional, reliable, and ever-improving system it is when parties of all stripes cooperate with each other.

Hence, regulatory bodies and governments as a whole can best serve the interests of their citizens by declining to take a superior position in disputes about how best to advance the Internet.

The Title II order took the U. S. and the Internet in the wrong direction. This is apparent from the delayed rate of improvement in Internet performance and continued complaints about the direction in which the Internet is headed.[13]

The FCC’s first instinct when it encounters a legitimate issue with Internet management should be to involve the multi-stakeholder community through such means as reaching out to the Broadband Internet Technical Advisory Group (BITAG), the Internet Engineering Task Force (IETF), the Internet Society, and professional organizations such as ACM and IEEE.

[1] Lawrence E. Strickling, “Moving Together Beyond Dubai,” NTIA, April 2, 2013,

[2] Federal Communications Commission, “Memorandum Report and Order in the Matter of Formal Complaint of Free Press and Public Knowledge Against Comcast Corporation for Secretly Degrading Peer-to-Peer Applications et Al.,” August 1, 2008,

[3] Susan P Crawford, “Comcast Is Pretending to Be You,” Susan Crawford, October 19, 2007,

[4] Matthew Lasar, “Cable and Telcos Side with Comcast in FCC BitTorrent Dispute,” Ars Technica, February 19, 2008,; Marvin Ammori et al., “Formal Complaint of Free Press and Public Knowledge Against Comcast Corporation For Secretly Degrading Peer-to-Peer Applications” (Federal Communications Commission, November 1, 2007),…/fp_pk_comcast_complaint.pdf.

[5] Personal communication with DOCSIS 3 standards developers.

[6] Jim Mills et al., “Comcast’s Protocol-Agnostic Congestion Management System,” n.d.,

[7] Wikipedia, “Bufferbloat,” Wikipedia, accessed July 15, 2011,

[8] Mirja Kuehlewind et al., “Low Extra Delay Background Transport (LEDBAT),” n.d.,

[9] Kathleen Nichols and Van Vacobson, “Controlling Queue Delay,” ACM Queue, May 6, 2012,

[10] Greg White, “How DOCSIS 3.1 Reduces Latency with Active Queue Management,” CableLabs, June 6, 2014,

[11] Edward Wyatt, “Court Favors Comcast in F.C.C. ‘Net Neutrality’ Ruling,” The New York Times, April 6, 2010, sec. Technology,

[12] Free Press, “Net Neutrality Violations: A Brief History,” Free Press, n.d.,

[13] Richard Bennett, “Open Internet Orders Degrade Internet Improvement,” High Tech Forum, June 19, 2017,; Nilay Patel, “The Internet Is Fucked (Again) – The Verge,” The Verge, July 12, 2017,