March 4, 2015 12:42 PMpress release describing its decision to reclassify broadband Internet access is a little different on the subject of interconnection than Chairman Wheeler's "Fact Sheet" 3 weeks earlier. The FCC's press release from last week is substantially similar to the Chairman's fact sheet, except that it contains no reference to the classification of the "service that broadband providers make available to 'edge providers.'"
The Chairman's "Edge Service" Classification Proposal Was Really Unpopular
On the last day that parties could lobby the Commission, (February 19th) Google had meetings with senior advisors for the Chairman and the two Democratic Commissioners. In these meetings, Google persuasively argued that "the Commission should not attempt to classify a "service that broadband providers make available to 'edge providers,'" because "this supposed additional service does not exist." Ex Parte (Internal quote to Chairman's "fact sheet.") On the eve of the Commission's vote, the Wall Street Journal reported that "the FCC tweaked its language to address Google's concern."
Without this report, it would be hard to know who to credit for killing a truly reckless idea. That's because the following 5 parties all made arguments against the Chairman's proposed "new service" on the same day as Google: Akamai, Free Press and New America's Open Technology Institute, National Hispanic Media Coalition, and Cox Communications.
When you somehow manage to get Free Press, New America, and the National Hispanic Media Coalition to oppose an additional Title II regulatory classification--because it makes your original Title II reclassification look even more legally suspect--that's when you know you've gone too far.
The Importance of Direct Interconnection Agreements
The scariest aspect of the Commission's proposal must have been the FCC's casual willingness to disturb these companies' existing arrangements with ISPs by mandating the future terms on which they and others would be able to obtain interconnection. For companies with little to no regard for the terms of their existing interconnection agreements, like Netflix and its transit vendors, the Commission could not make their situation worse. But, for the leading Internet companies, the Commission must have appeared alarmingly ignorant of/indifferent to the importance of these agreements to Internet traffic delivery.
Reason 1: The Disintermediation of Internet Traffic
In 2009, the University of Michigan, Arbor Networks, and Merit Network presented the results of the largest traffic study since the advent of the commercial Internet. The study showed that, between 2007 and 2009, Internet traffic delivery changed radically. Over only 2 years, Internet transit (the traditional "intermediary" between ISPs) became dramatically less important as a traffic delivery vehicle.
Instead, major content providers began delivering more and more content directly to the consumer's ISP (either through their own networks or CDNs). Accordingly, since 2009, the "tech giants" have been accelerating their investment in network assets and data centers to route their high bandwidth traffic directly to efficient delivery points in the ISP's networks.
Reason 2: Interconnection Agreements Reflect Valuable Investment
As the Internet has evolved to more efficiently deliver high-bandwidth content to consumers, the largest content providers--including Netflix for the first 4 years of its streaming service--have placed a premium on placing content closer to the customer. Therefore, the largest traffic sources have entered into agreements to directly exchange traffic with their customers' ISPs. When these agreements provide for the settlement-free exchange of traffic, it is because this reflects the mutual benefits received by both parties.
Since parties to settlement-free "peering" arrangements each provide the other with valuable network facilities, or other benefits, this value can be observed by looking at the investment the parties put into their networks (i.e., capital expenditures). To get an idea of the importance of those agreements to Internet companies, consider the increase in capex by the largest Internet companies since 2009.
Regulation of Interconnection Terms Could Devalue Previous and Future Investment
As we can see from the chart above the major Internet companies have undertaken a massive amount of capital spending over the past 6 years in order to efficiently deliver content and services to consumers. To be sure, not all of the Internet companies' capex is driven by traffic delivery interconnection concerns, but the increase in these companies' capex since 2009 correlates with the findings of the University of Michigan, et al., traffic study referenced above. Moreover, news reports have confirmed spending on improved data networking infrastructure as a capex driver. See, e.g., here and here.
This capital investment has been made by edge companies and CDNs with the expectation that it will allow these firms to provide a better experience to their customers than their competitors provide. Indeed, Google notes that it has entered into peering agreements with some of the largest ISPs because it is "unable to use transit to reach users on those networks with reasonable quality." Ex parte at 2 (emphasis added.)
The risk of requiring ISPs to provide interconnection as a separate common carrier service was articulated succinctly by Akamai, which handles 15-30% of the world's Internet traffic. Akamai argued that the FCC must not mandate the terms and conditions of ISP interconnection, because if the ISPs are required to provide access on equal terms to all:
This is not technically feasible and the result could be access for none, which would decrease the performance, scalability, reliability and security of the Internet.Akamai, February 20th ex parte at 1 (emphasis added). In other words, Akamai understands and accepts that it "must often compete with others for access to ISP facilities." Akamai, February 9th ex parte, at 3 (emphasis added). But, does the FCC accept interconnection as a legitimate element of competition?
The Commission Should Not Displace Competition with Regulation
Netflix, Cogent, and Level 3 assert that they cannot get interconnection with the ISPs--on the terms they would prefer--because of a lack of competition. But, as Akamai explained, companies seeking the most efficient terms of distribution to the ISP's customers are competing with each other for the best access to these customers. Could it be that competition is the reason the transit companies aren't getting the terms they want?
Compare the sum of the capex of Netflix's distribution chain over the relevant time period, with their competitors. Is it surprising that Internet transit wants the FCC to "level the playing field?"
Transit is at a disadvantage relative to direct interconnection because the Internet has evolved. For the content distributors sending the most traffic, transit has not been the preferred solution for a long time. But competition, and not a lack thereof, produced this outcome.
Unlike the major focus of the larger net neutrality debate--which is concerned with adopting rules to foreclose future ISP service offerings--the regulation of interconnection terms and conditions is fraught with risks to existing service configurations. The FCC must be careful not to use prescriptive regulation of ISP interconnection terms--in the name of competitive "neutrality"--to foreclose innovative firms (and their customers) from reaping the benefits of their own ideas, risks, and investments.