Any decision about FCC regulation of the broadband Internet must account for the sunset of the public switched telephone network (PSTN) and the evolution of the Internet since the FCC issued its Cable Modem Order in 2002. The PSTN sunset and the design features of next generation networks have significant implications for Title II reclassification.
Reclassification Would Require Revisiting the Communications Act’s Core Definitional Conundrum
Title II regulates communications “services”, not particular types of infrastructure. For example, the definition of “telecommunications service” applies “regardless of the facilities used.” See 47 USC § 153(53).
The definitions of “information” and “telecommunications” services adopted in the 1996 Act track the distinction between “basic” and “enhanced” services that was adopted by the FCC in the Computer II proceeding’s Final Decision. See Verizon v. FCC, 740 F.3d 623, 630 (DC Cir. 2014).
In a post-PSTN world, there will no longer be any “basic” service as it was understood in the Computer II era or “telecommunications” as that term is defined in the 1996 Act.
Title II reclassification would thus require the FCC to revisit the core definitional conundrum of the Communications Act. The likely result is that many services that are currently unregulated would be subject to Title II regulation, including services that were once considered part of the Internet’s “edge”.
The Current Distinction Between Telecommunications and Information Services is Premised on the Existence of the PSTN
Both “telecommunications” and “information services” undoubtedly provide “communications” capabilities. “Information” services are nevertheless exempted from Title II regulation.
As a practical matter, the FCC uses a simple test to distinguish between communications services that are regulated under Title II and services that are not.
The FCC generally considers a service to be “telecommunications” if is “interconnected” with the PSTN — i.e., it is capable of communicating with all other devices that are connected to the PSTN — or has historically been subject to Title II regulation before the broadband era. With limited exceptions (e.g., Ethernet), services that are not interconnected with the PSTN are considered “information” services.
The FCC has generally relied on this simple test without reference to the specific language in the definitional section of the Communications Act. For example, although the FCC has refused to classify VoIP services for over a decade, it has imposed a multitude of Title II obligations on “interconnected” VoIP services because they are interconnected with the PSTN (and thus substitutable with POTS).
Most definitions in FCC regulations are also premised on the PSTN as well. For example, the term “special access” is defined as access that bypasses the local switch. When the PSTN sunsets, there won’t be local switches, which will render meaningless the current definition of “special access”.
Any attempt to define a new line between regulated and unregulated communications services would require the FCC to revisit the conundrums of the Computer II proceedings and the statutory language defining “telecommunications” and “information” services.
Definitional Lessons Learned From the Computer II Proceedings
The current distinction between information and telecommunications services is derived from the Final Decision of the Computer II proceedings. The FCC adopted a relatively clear line in the Final Decision because its earlier, more nuanced definitions had proven unworkable.
The FCC’s early efforts to distinguish between basic and enhanced services recognized that the primary purpose of some services that employ “computer processing” is to offer communications capabilities that are substitutable for the traditional PSTN services subject to regulation under Title II. See Regulatory and Policy Problems Presented by the Interdependence of Computer and Communication Services and Facilities, Tentative Decision of the Commission, 28 FCC 2d 291 at ¶¶ 41-42 (1970).
The FCC thus attempted to distinguish between “enhanced” services that were primarily intended to provide communications, which were subject to regulation, and “enhanced” services that were not primarily intended to provide communications, which were not subject to regulation. This resulted in regulatory uncertainty that prompted companies to ask the FCC for permission before offering innovative services.
In its Final Decision, the FCC reversed course. It concluded that any attempt to “delineate a distinction between communications and data processing services . . . would be ultimately futile, inconsistent with our statutory mandate, and contrary to the public interest.” Second Computer Inquiry, Final Decision, FCC 80-189 at ¶ 107 (1980).
The reasons provided by the FCC then are still applicable today.
As the market applications of computer technology continue to evolve we believe that attempts to distinguish enhanced services either will fail or result in an unpredictable and inconsistent scheme of regulation. This is because a definitional structure is not independent of advances in computer technology and its concomitant market applications. A certain degree of flexibility must be maintained to accommodate these advances. To the extent flexibility is incorporated into the definitions, there is a corresponding degree of uncertainty. Thus the boundary line differentiating enhanced communications and data processing services can vacillate, and confidence in decisions made based on that distinction would be diminished. Final Decision at ¶ 107.
Computer technology is increasingly removing technical limitations as to the types of enhanced services that may be offered. Yet, a classification scheme which would categorize enhanced services as either communications or data processing inherently limits the types of services that an unregulated entity may offer. Providers of data processing and other computer services acquire the necessary transmission facilities from communications common carriers pursuant to tariff, and resell this transmission capability as part of their enhanced offering. At the same time, an entity which acquires the same transmission facilities from a carrier and offers a ‘communication’ service is presently regulated as a common carrier under Title II of the act. Accordingly, a resale entity is regulated as a common carrier only if it is providing a communications service. Final Decision at ¶¶ 107-108.
The FCC found that the difficulty of distinguishing between services that involve data processing and resellers of telecommunications services would restrict output.
First, the vendor of unregulated enhanced services may not provide an enhanced ‘communications’ service. This means that its services must be artificially structured so as to not come under our regulatory umbrella (the guidelines for which we have already concluded would be less than precise). To the extent services must be so structured there is a corresponding inability to fully tailor services to consumer needs. This has the result of artificially restricting the supply of services provided over the telecommunications network. Final Decision at ¶ 111.
Google’s decision not to offer a bundled voice service with Google Fiber in Kansas City is a contemporary example of the ways in which Title II regulations restrict output.
The FCC also noted that its tentative definitional distinction would reduced innovation by subjecting previously unregulated enhanced service providers to Title II regulation.
The second public interest implication is the increased potential for expansion of regulation over currently unregulated providers of information or data processing services. . . . A resale carrier could offer any enhanced service, whereas the unregulated vendor of computer services may offer only those enhanced services that are not regulated as ‘communications.’ . . . One result may be an indirect forcing of currently unregulated entities to acquire common carrier status in order to obtain the same degree of flexibility afforded a resale common carrier.’ . . . In addition to increasing the scope of Commission regulation, the specter of potential regulation may impose artificial barriers to entry. Final Decision at ¶ 112.
For these reasons, the FCC concluded that “all enhanced computer services should be accorded the same regulatory treatment and that no regulatory scheme could be adopted which would rationally distinguish and classify enhanced services as either communications or data processing.” Final Decision at ¶ 113.
This left the FCC with two options: (1) “subject all enhanced services to regulation, or  refrain from regulating them in toto.” Final Decision at ¶ 114.
The FCC adopted the second option and concluded that enhanced services should not be regulated under Title II. “In defining the difference between basic and enhanced services, we have concluded that basic transmission services are traditional common carrier communications services and that enhanced services are not.” Final Decision at ¶¶ 119-20. This conclusion created the current test, which regulates services that are interconnected with the PSTN or that were historically regulated in the pre-broadband era.
Telecommunications and Information Services Are Inextricably Intertwined in an All-IP World
Once the PSTN sunsets, the definitional dichotomy in the Computer II proceeding’s Final Decision will no longer provide a simple demarcation point between regulated and unregulated services.
In an all-IP world, no communications will be interconnected with the PSTN. In the terms of the Final Decision, all “communications and data processing technologies [will] have become intertwined so thoroughly as to produce a form different from any explicitly recognized in the Communications Act.” See Final Decision at ¶ 120.
POTS switches will be replaced entirely by internetworked systems of servers, routers, links, and devices that are owned and operated both by end users and a multitude of service providers who offer a multitude of different services at wholesale and to businesses and consumers.
There Are No Separable “Telecommunications” Components in an All-IP World
The evolution of the Internet renders it impractical and undesirable for the FCC to separate only certain aspects of “last mile” connections to the Internet from other aspects of broadband Internet access on a definitional basis.
Even if the FCC were to adopt an infrastructure-based definition of “telecommunications”, it would by necessity include servers, routers, transit, and end user devices within its scope.
Today, there are no separable portions of the infrastructure underlying broadband Internet access that could meet the 1996 Act’s definition of “telecommunications”.
In the past, the FCC has indicated that broadband Internet access uses an underlying “telecommunications” component to provide information services.
‘[a]ll information services require the use of telecommunications to connect customers to the computers or other processors that are capable of generating, storing, or manipulating information.’ Although the transmission of information to and from these computers may constitute ‘telecommunications,’ that transmission is not necessarily a separate “telecommunications service.” We are not aware of any cable modem service provider that has made a stand-alone offering of transmission for a fee directly to the public, or to such classes of users as to be effectively available directly to the public. Cable Modem Order, FCC 02-77 at ¶ 40 (2002).
The FCC didn’t directly address whether broadband Internet access uses “telecommunications” as defined by the 1996 Act because, “For the Commission . . . the question whether cable broadband Internet providers ‘offer’ telecommunications involved more than whether telecommunications was one necessary component of cable modem service.” Brand X, 545 U.S. 967, 988 (2005).
The Court gave Chevron deference to the FCC’s determination that the underlying facilities used to provide broadband Internet access are not a “telecommunications service” that is separable from the “information service” because, “As provided to the end user the telecommunications is part and parcel of cable modem service and is integral to its other capabilities.” Cable Modem Order at ¶ 39.
As Defined in the 1996 Act, “Telecommunications” Does Not Exist Separately from a Broadband Internet Access Offering
Even if it were reasonable to conclude that “telecommunications” was “part and parcel” of broadband Internet access in 2002, it would not be reasonable today.
The 1996 Act defines “telecommunications” as “the transmission, between or among points specified by the user, of information of the user’s choosing, without change in the form or content of the information as sent and received.”
This definition was created for the dial-up Internet, in which end-users dialed a telephone number to create a circuit switched connection and used modems to send audible tones (conceptually similar to Morse code) through PSTN switches. (The word “modem” is short for modulation and demodulation of these audible tones).
The “net neutrality” concept of “edge” providers operating only at higher “layers” was based on the fact that the dial-up Internet did not require fundamental changes to the circuit-switched architecture of the PSTN.
The broadband Internet doesn’t work that way. Users of broadband Internet access services (1) don’t “specify” any points of transmission by dialing a telephone number, (2) don’t “choose” the information that is transmitted and received as a result of typing in a URL, and (3) the information sent and received is routinely changed in form and content without the user’s knowledge.
Broadband Users Don’t Specify Points of Transmission
On the POTS network, a user specified a point of transmission by dialing a telephone number to reach a particular local exchange point. On the broadband Internet, a user who types in a particular URL is not specifying a transmission path with any defined points. The information that is transmitted as a result of the user’s entering a URL may be delivered from any of a number of potential locations using any number of potential routes.
Broadband Users Don’t “Choose” the Information Transmitted and the Form and Content of Broadband Transmissions Changes Routinely
When a user types in a URL, an elaborate set of information transmissions typically occurs without the user’s knowledge.
For example, in addition to the information associated with a URL that a user chooses to type in, a user who is web browsing often receives ads from third-party ad servers that the user (and, increasingly, the owner of the accessed webpage) did not choose. Though it appears to the user that the webpage is an integrated whole, the ads are typically received from a different server in a different location from the information associated with the URL (which itself may be served from multiple “points”). This is an inherent feature of contextual advertising on the Internet.
The SPDY protocol — a potential replacement for HTTP — is another example of the ways in which the modern Internet does not fit the definition of “telecommunications”. Google’s Chromium Project notes that, “the web pages transmitted today are significantly different from web pages 10 years ago and demand improvements to HTTP that could not have been anticipated when HTTP was developed.” (Emphasis added) One of the ways that SPDY improves latency is by allowing servers to push data to a client before the client asks for it — a feature that isn’t contemplated by the Act’s PSTN-based definition of “telecommunications”.
Problems Caused by Attempting to Separate “Telecommunications” from “Information Services” in an All-IP World
Net neutrality is premised on the notion that “last mile” connections to the Internet are uniquely problematic, but the definitions in the Act are not limited to particular facilities or unique circumstances. Distinguishing “telecommunications” from “information services” based on infrastructure ownership would expand the scope of Title II to entities that have never before been considered common carriers.
In the proceeding to classify cable broadband as an “information” service, those who favored regulation of cable broadband claimed that “the Communications Act unambiguously classifies as telecommunications carriers all entities that use telecommunications inputs to provide information service.” Brand X at 994. In Brand X, the Court noted that “this argument would subject to mandatory common carrier regulation all information-service providers that use telecommunications as an input to provide information services to the public.” Brand X at 994. That would include all forms of VoIP, search engines, devices, and applications.
For example, the FCC would likely receive Chevron deference if it were to subject mobile operating systems and the app stores operated by Google and Apple to Title II regulation under an infrastructure-based definition of “telecommunications”. In the Cable Modem Order, the FCC recognized that, “when a cable modem service subscriber initiates his cable modem service, the cable modem service subscriber’s computer becomes a part of the Internet.” Cable Modem Order at n. 55. Mobile handsets and operating systems are just as necessary to provide broadband Internet access as the rest of the wireless infrastructure, and the vendors of handsets and mobile operating systems present similar competitive and “social contract” issues as wireless carriers.
A recent example is the Mobile CPNI Order, in which the FCC declared that mobile wireless carriers must protect customer proprietary network information (CPNI) that carriers cause to be stored on their customer’s mobile devices, FCC 13-89 at ¶ 5 (2013), even though Section 222 of the Communications Act imposes a duty to protect CPNI only on “telecommunications carriers”. See 47 USC § 222(a) (this provision had already been extended to “interconnected VoIP” providers under the FCC’s ancillary authority (VoIP CPNI Order, FCC 07-22 at ¶¶ 54-55 (2007)). Mobile OS and app store providers have access to the same types of CPNI the FCC currently requires wireless carriers to protect. For example, iOS provides users with options to automatically send network and usage data to Apple and to automatically update applications purchased through Apple’s App Store. Another example is the IP Multimedia Subsystem (IMS), which provides mobile operating systems and applications with access to CPNI stored in a carrier’s IMS core network and allows a carrier to provide OSs and apps with direct API-level access to CPNI in the IMS core.
Any attempt by the FCC to define an infrastructure-based regulatory demarcation for the modern Internet would suffer from the same defects as the agency’s initial attempts to define regulated and unregulated “enhanced” services. Attempting to define prophylactic “lines in the sand” would limit, rather than promote, innovation by forcing innovators to ask the FCC for permission.