Last week, the Federal Communications Commission confirmed that it will not expand the scope of the ongoing net neutrality proceeding to encompass peering and transit services. This was a blow to Netflix CEO Reed Hastings, who launched a blistering attack last month designed to focus the agency’s attention on this market. But given the robust state of competition in the interconnection market, the FCC’s response was good news for the future of the Internet ecosystem.
The brouhaha began with last month’s announcement that Comcast and Netflix had reached a “mutually beneficial interconnection agreement.” Though the companies did not disclose the terms of the deal, most believe that Netflix has agreed to pay to connect its servers to the Comcast network and stream content directly to Comcast consumers. Columbia Law Professor and net neutrality guru Tim Wu decried the deal as “water in the basement of the Internet industry,” suggesting this was the beginning of the end of the Internet as we know it.
This doom-and-gloom stems in part from a misconception that Netflix generally pays nothing to upload its content to the Internet. This assumption is false, as I explain at length in my most recent Perspectives article for the Free State Foundation. In fact, Netflix and other content providers contract with transit providers such as Cogent and Level 3 Communications. These transit providers then enter into agreements to exchange traffic with other networks, including broadband providers such as Comcast and Verizon. This is the Internet as we know it: a largely unregulated network of networks, governed by privately negotiated interconnection agreements whose terms can vary dramatically based upon the needs of the specific signatories to any given agreement.
Moreover, despite Hastings’ hand-wringing, the Comcast-Netflix agreement illustrates the robust competitiveness of the interconnection market. Beginning in 2013, Netflix customers began complaining about declines in service quality when streaming over Comcast and Verizon networks. The blame appears to lie with Cogent Communications, one of Netflix’s primary transit providers. Netflix has grown large enough that its traffic has overwhelmed the existing connections through which data travels between the Cogent and Verizon networks. The two companies have been locked in a year-long battle over the cost of upgrading those connections to allow greater volumes of traffic. In the meantime, Netflix customers have suffered, as it is primarily Verizon-bound Netflix traffic that suffers from congestion at these connection bottlenecks. It is likely that a similar dispute is responsible for delays in Comcast-bound Netflix traffic. Tired of bearing the cost of this dispute, Netflix elected to take its transit business elsewhere – signing an interconnection agreement with Comcast directly, rather than relying upon Cogent for delivery of its Comcast-bound traffic. In essence, Netflix eliminated the middleman – and it was competition in the transit market that allowed it to escape its position as a hostage in a battle between network providers.
Indeed, pulling back the curtain on the interconnection market reveals a rich and diverse competitive landscape. Content providers can choose from a wide range of transit providers, some of which specialize in transit alone, while others provide a variety of complementary services as well. Some provide service only within a limited geographic area, while others operate across continents. Transit providers famously subsist on razor-thin margins. Transit prices have fallen, dramatically, every quarter and will continue to do so for the near future.
In the interests of “strong net neutrality,” Hastings would replace these rich and diverse business agreements with a uniform duty to carry all content providers’ traffic at zero cost. It is not at all clear how such a market would work. Would Netflix still pay transit fees, as it does now? If so, to whom? In a YouTube age where many end users are also content providers, who would get to partake of this special zero-cost rate? And who would be left holding the bag to pay the costs of network operations?
These questions show that Hastings’ latest complaint is simply part of his long-standing quixotic quest to reduce Netflix’s transit costs – costs that, as co-blogger Bret Swanson explains, are an input for Netflix the same way steel is for GM – to zero. Over the past few years, the company built its own content delivery network, known as OpenConnect, as an in-house alternative to Cogent and other transit providers. Netflix tried to entice broadband providers to exchange traffic with OpenConnect for free, in essence reducing its transit costs to zero. And although some providers agreed, most major broadband providers did not. Having failed to negotiate his favored price in the marketplace, Hastings is now trying to force his price on the market by regulatory fiat – a move that anyone with an elementary exposure to public choice theory would both anticipate and condemn.
There are good economic reasons why the market price for Netflix traffic is above zero, as co-blogger Babette Boliek notes. Netflix consumes a substantial portion of the nation’s broadband networks – it is responsible for almost one-third of all North American Internet traffic during peak hours. Operating those networks is not costless; they are the result of literally hundreds of billions of dollars of sunk investment, and several billion dollars more each year in growth, maintenance, and upgrade costs (including upgrading networks to process growing volumes of Netflix traffic). Given that Netflix benefits from ongoing network operations, it is not surprising that the market demands that it also contribute to the costs of deployment and upkeep.
The fact that Netflix pays Cogent transit fees is uncontroversial. The Netflix-Comcast agreement should be no less controversial merely because the identity of the transit provider has changed, or because the transit provider also happens to manage a last-mile broadband network. The FCC is absolutely right that the market for interconnection is robust and competitive, meaning there is no reason to justify regulatory intervention. Its regulatory humility is both refreshing and promising for the future of Internet policy.
Download the full Perspectives article for a more in-depth discussion of the Comcast-Netflix agreement.