1 Gillan Associates 1.0 Introduction Throughout the world, the technology used to provide wireline phone service is changing from a circuit-switched architecture to a technology based on the same suite of protocols (IP) that support the Internet. In most nations, this technological transition presents little regulatory confusion because there are no core questions of regulatory authority or policy tied to the change in technology. In the United States, however, the transition is complicated because there is a statutory distinction between information and telecommunication services, and some have argued that the change in technology to IP changes the nature of the service. The purpose of this paper is to describe the technological drivers of the transition and, to better understand the complexity of the US policy response, to explain (presumably to a European audience) 2 the partial paralysis this unique statutory construct has caused to the transition in the United States. 2.0 The Central Technical Elements of the Transition 3 The transformation of public phone networks to IP technology is an architectural response to the ascendancy of data communications. The traditional telephone network was designed to support voice service, 4 with data services provided as an overlay. With 1 is a consulting economist with a practice specializing in the competitive implications of different policy choices in the telecommunications industry. Although not a lawyer, telecommunications policy in the United States (as elsewhere) is unavoidably the product of laws and rules that reflect economic choices as well as legal constraints. As such, it is not possible to describe the policy landscape in telecommunications without including indeed, as shown here, without first starting with a legal background to frame the issue. Nothing in this paper, however, should be interpreted as legal opinion. 2 This paper was prepared for the 6th Annual Competition in Regulated Industries Conference, November 22, 2013, Brussels, Belgium. 3 Portions of this section previously appeared in similar form in a paper (jointly prepared by the author) that focused on one element of the transition, specifically the interconnection of rival IP networks. See The Transition to an All-IP Network: A Primer on the Architectural Components of IP Interconnection, and David Malfara, National Regulatory Research Institute, May 2012, NRRI 12 05. 4 The existing network architecture is frequently characterized by two defining characteristics: (1) its reliance on time-division multiplexing (TDM) to derive individual channels on a common transmission facility (be it copper or fiber), and (2) the use of circuit-switching to establish temporary connections between TDM channels on different facilities. Circuit-switched
IP, these roles are reversed, with the underlying physical network designed for data service and voice is provisioned as a specialized data stream (with priority) to ensure quality. To be sure, the technology of the wireline phone network has changed before. It has moved from its roots as an analog network with in-band signaling to its current architecture characterized by high-speed fiber transport, digital switching, and call control managed by a parallel signaling network, Signaling System Seven (SS7). These evolutions, however, were largely silent that is, the changes were known primarily to those who managed (or regulated) the network, but were generally invisible to consumers (and the political process). 5 The emergence of IP technology is far more visible. One explanation for the visibility of IP technology is that it supports one of the most dynamic innovations of our time, the global Internet. Internet Protocol is used in networks other than the Internet, however, and the importance of IP technology extends beyond this one (albeit remarkable) network. The focus here is on the use of IP technology as the core technology underlying the public phone networks of the future, and the particular visibility the technology has achieved in the United States, residing as it does at the center of a contentious political and policy debate. 2.1 Achieving Quality of Service (QoS) in an IP-based Phone Network A key feature of today's wireline phone networks is the expected quality of voice service. As these phone networks transition from a circuit-switched to a packet-switched architecture based on IP, a key question is how quality will be maintained. The quality dimension that is most important to understanding the transition to IP networks is dealing with congestion, and understanding how congestion will be addressed in these new networks. 6 networks reserve a dedicated channel for the duration of a call, while packet networks constantly (and rapidly) break the call into individual packets that are transmitted individually on a shared channel and then reassembled at the far end. In this paper, circuit-switched and TDM are used interchangeably to describe the traditional architecture. 5 This is not to say that prior technological changes did not affect consumers through improvements in quality (such as the significant post-dial delays eliminated by SS7), reductions in cost, or changes in competitive conditions. The point rather is that the technology itself did not become the source of different policies or consumer perceptions. 6 Many factors affect the perceived (by an end user) quality of a phone call, including noise and the audio frequency range being transmitted. Because both traditional and packet networks are digital, noise issues are generally resolved. As to the available audio frequencies, this parameter is limited in circuit-switched networks by the network's design and standardization. Because packet networks can flexibly assign bandwidth to different applications, IP networks are capable of high-definition voice service (HD voice) that increase the fidelity of the audio stream and thus improve sound quality. The quality metric most impacted by congestion, 2
In a circuit-switched network, an external signaling network (SS7) addresses the potential for congestion directly at its source, when a call is first dialed. SS7 immediately determines whether a circuit is available all the way through to the called number. If so, the call proceeds; if not, the caller is sent a busy signal. In other words, SS7 addresses congestion through avoidance, by blocking the congestion-producing call (and only the congestion-producing call). The public Internet addresses congestion with a very different philosophy. The public Internet is referred to as a best-efforts network because it is designed to treat all packets the same way. In a best-efforts architecture, queuing is used to address congestion. The network delays some packets (i.e., it buffers the traffic ), either at its source or at a router, while other packets are transmitted. 7 The result is that all packets are created equal and, therefore, are equally likely to be delayed. Because the process of buffering (or discarding) packets can be arbitrary, packets from all information flows are likely to be affected by packet loss and degradation during a period of congestion. On the Internet, therefore, every user can experience the consequences of congestion. In contrast, a circuit-switched architecture addresses congestion by blocking some calls entirely, but otherwise maintains the quality of the calls that do go through. The key to successfully replacing traditional technology with an IP-based infrastructure is to find a way to prioritize the data-streams associated with a voice call so as to emulate the performance of a circuit-switched architecture in transmitting in realtime (i.e., without delay) voice packets. 2.2 Managed IP Networks vs. Over-the-Top Applications The single term VoIP (i.e., Voice over Internet Protocol) is commonly used to describe two very different approaches to using IP technology to provide voice service. The first is VoIP over a managed network; the second configuration is VoIP as an undifferentiated Internet application. Unfortunately, there is not complete consensus over the terminologies used to describe this distinction. 8 however, is the packet-speed necessary to maintain the existing quality that consumers experience with circuit-switched networks. 7 If congestion becomes too much of a problem, the network may even discard some of the "overflow" packets. 8 Canada has the most intuitively appealing terms, using the term access-dependent VoIP to refer to a service that is controlled by the network provider serving a customer, and access independent VoIP to refer to an application that merely rides the public Internet without any specialized treatment. In the United States, the term nomadic or over-the-top is frequently used to describe an access-independent application (recognizing that the equipment that makes possible the call is frequently connected by the customer and can be moved to a different Internet connection without loss in functionality), and fixed VoIP is used to describe services that are 3
The term used here to describe the first approach is a managed-packet VoIP service, which is designed to achieve traditional quality while preserving the flexibility of a packet-based network. A managed-packet network is an IP network that is capable of defining class-of-service categories and then adapting network performance to maintain a quality-of-service level appropriate to each class. These networks are generally known as next-generation networks (NGNs), 9 and their importance stems from their ability to resolve congestion and thereby ensure service quality. Managed networks are particularly useful for a real-time service such as VoIP, which requires that packets be delivered reliably and without significant delay or variance (latency or jitter). Managed-packet networks are a key improvement over a best-efforts IP network such as the public Internet. A managed-packet network can provide the higher priority to voice needed to achieve the same quality as a circuit-switched network. In contrast, an over-the-top VoIP service (such as Vonage or Skype) rides the public Internet for some part of the service and these packets are treated no differently (during this part of their transmission) from other Internet traffic. In circumstances of low congestion, over-thetop VoIP services can provide quality equivalent to that of managed networks, although that quality cannot be guaranteed because the service relies on the best-efforts design of the Internet. In the United States (as elsewhere) most providers of VoIP separate the voice packets from other traffic on the Internet in order to preserve quality. This is true, for instance, for incumbent local exchange carriers such as AT&T (U-verse) and Verizon (FiOS), as well as cable-based VoIP providers and other competitors (that primarily serve the business market). Overall, it is estimated that approximately 90% of the VoIP traffic in the United States is of the managed (or access-dependent) form. 10 Historically, new technologies have been introduced into the United States public telephone networks without significant controversy, and certainly without substantial debate as to whether regulatory policies would continue to apply. Rather, the general approach has been for the provider to define its service, but that the choice of technology remains with that provider without limit. 11 In this instance, however, the transition to IP technology has provoked a policy debate as to whether the change in technology fundamentally upsets the regulatory framework applicable to (what has always been considered) phone service in the United States. dependent upon the provider of the broadband connection (and therefore controls the information flow to achieve desired quality). 9 10 http://www.itu.int/itut/studygroups/com13/ngn2004/working_definition.html Local Telephone Competition: Status as of June 30, 2012, Industry Analysis and Technology Division, Wireline Competition Bureau, Federal Communications Commission, June 2013, Figure 5 at 7. 11 In instances where a performance characteristic is affected by a particular network technology, the difference is typically identified in the tariff setting forth the terms and conditions of service, but would not become a choice-variable of the customer. 4
3.0 The Distinction between an Information Service and Telecommunication Service in US Law Having established the basic parameters of the transition of the phone network to packet technology (which is a world-wide development), it is important to understand the peculiarities of US law, in particular the distinction between information services and telecommunication services and the consequences that (have been claimed to) follow from that distinction. First, the basics the definitions themselves: INFORMATION SERVICE.--The term ''information service'' means the offering of a capability for generating, acquiring, storing, transforming, processing, retrieving, utilizing, or making available information via telecommunications, and includes electronic publishing, but does not include any use of any such capability for the management, control, or operation of a telecommunications system or the management of a telecommunications service. 12 TELECOMMUNICATIONS.--The term ''telecommunications'' means 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. 13 To the untrained eye, these definitions would seem mutually exclusive and grounded in the practical difference between transmitting information without changing its content (like the information that replicates the human voice in a phone call), and a service that enables the end-user to manipulate (i.e., store, transform or retrieve) information in a way that creates value. As explained below, however, it is not that these definitions could not be clearly understood and applied, but that the consequences associated with the definition have caused the federal regulatory authority (the Federal Communications Commission or FCC) 14 to largely avoid determining the regulatory classification of VoIP, thereby creating a regulatory vacuum and incessant debate. 12 13 14 47 U.S.C. 153 (20) 47 U.S.C. 153 (43) In addition to the Federal Communications Commission, each state typically has a state regulatory agency that has its own jurisdiction over some retail services, and shared jurisdiction over wholesale (i.e., carrier-to-carrier) relationships. The complex interrelationship between the federal and state jurisdictions in the United States is beyond the limited goals of this paper. 5
3.1 The Regulatory Implications of Classification Managing the Transition to IP-Based Generally speaking, the federal Communications Act regulates telecommunications carriers but not information-service providers as common carriers. Among other duties, telecommunications carriers must charge just and reasonable, nondiscriminatory rates to their customers, and design their networks so that other carriers can interconnect with them. Moreover, if a local exchange carrier, the Act establishes clear obligations to exchange traffic and further, if an incumbent local exchange carrier, additional provisions requiring unbundling of certain network facilities, as well as a number of important requirements addressing interconnection. Information-service providers, by contrast, are not subject to common-carrier regulation, although the FCC has the ability to impose some regulatory obligations under its ancillary jurisdiction. The boundaries of this additional authority are currently subject to uncertainty in the US judicial system, but the important point for the purposes of this paper concerns the (potential) interplay between these definitions and the core consumer and competitive protections that apply to telecommunications services. 3.2 A History of Uncertainty For a number of reasons, the Federal Communications Commission has chosen to incrementally subject VoIP to a variety of regulatory obligations without ever directly addressing the central question of the services classification. 15 Over time, the Commission has generally subjected VoIP services to the following obligations (while skirting the classification question): 16 Routing calls to E911 with location information; Payment of federal and state universal service obligations; Support lawful wiretaps; Protection of customer proprietary network information; Restrictions on the discontinuance of service; Disabilities access; Telephone number porting; Prohibitions on call blocking; Payment of intercarrier compensation obligations; and Reporting a variety of network performance statistics. 15 For the purposes of this paper, the reasons that the FCC has avoided answering the direct question as to whether VoIP is an information service or telecommunications service is not as relevant as the fact that it has avoided that determination. 16 Presentation of Glenn S. Richards, Executive Director, Voice on the Net Coalition, before the Washington Utilities and Transportation Commission, WA UTC Docket UT-131989 November 12, 2013. 6
Thus far, the FCC has only unambiguously decided the regulatory classification of two services that represent the polar ends of the spectrum of services provided using VoIP technology. First, anchoring on end of the spectrum, the FCC determined that Free World Dialup (FWD) was an information service. 17 This service allowed users to become members and talk to each other over the Internet. Members were required to have an existing broadband Internet access service, because FWD did not offer any transmission service or transmission capability. In addition, members were required to acquire and appropriately configure specialized (SIP) phones or download software that enabled their personal computers to function as soft phones. Once these criteria were met, members could obtain a FWD-assigned number (not a number associated with traditional phone service) to facilitate using the member s broadband service to establish peer-to-peer communications with other members. The members end-user devices (not FWD) established the actual connections and manage the calls, with the public Internet providing the transmission. Because of the particular features of this service, the FCC concluded that this service was an information service. At the other end of the spectrum, the FCC determined that an IP-in-the-middle service offered by AT&T was a telecommunication service. This decision addressed an AT&T service that relied upon conventional telephones at the end-points of the call, but at some point in the internal AT&T long distance network, the call was carried over the Internet. In this decision, the FCC concluded that the service was a telecommunications service, explaining that it (1) uses ordinary customer equipment with no enhanced functionality; (2) originates and terminates on the traditional public telephone network; and (3) undergoes no net protocol conversion and provides no enhanced functionality to the customer due to the provider s use of IP technology. 18 17 Memorandum Opinion and Order, Petition for Declaratory Ruling that Pulver.com s Free World Dialup is Neither Telecommunications Nor a Telecommunications Service, Federal Communications Commission, WC Docket No. 03-45, Rel. February 19, 2004. 18 Many IP-based services arguably satisfy the criteria used to define the AT&T IP-in-the- Middle as telecommunications (even though the AT&T decision was limited to long distance calling, while virtually all VoIP services today involve both local and long distance calling). To date, however, the FCC has not explicitly concluded that any other services satisfy the key prongs of the definition. It is not the purpose of this paper to explain why a particular service is or is not a telecommunication service if held against these federal standards. Rather, the point here is that the FCC itself has chosen to avoid the analysis, therefore creating a persistent uncertainty. 7
3.3 The Current State It is unclear whether being the subject of an FCC decision is equivalent to a curse, but it is worth noting that Free World Dialup appears to have ceased operation, 19 and AT&T now proudly advertises that its VoIP services never touch the Internet. 20 Either way, the services addressed by the FCC were never commercially significant (despite their legal prominence), and have both disappeared from the market. In contrast, over 90% of the VoIP services in the United States today fall in the gray area between these extremes, and the most fundamental regulatory question that is, are VoIP services telecommunications services or information services remains unanswered by the FCC. The FCC s avoidance of the core question has proven particularly unsettling to state public utility commissions that would have no jurisdiction should VoIP be declared an information service. This frustration was recently summarized by a leading state regulator (in his role representing all state regulators) before the United States Congress: For over 10 years, the agency s [FCC s] inability, under both Democratic and Republican Chairmen, to provide needed certainty by classifying VoIP services as either a telecommunications service or an information service has continued to undermine the telecommunications market and spawn a plethora of unnecessary agency and court proceedings. It [the FCC] has left this question unresolved for over 10 years. It has also left some consumers who choose IP-based services with fewer protections than users of the older circuit-switched/copper network have, even though, from the consumer perspective, the voice service offered is exactly the same. NARUC, the States, and industry stakeholders continue to waste significant resources, all at the ultimate expense of the taxpayer and ratepayers, on proceedings that would be unnecessary if the FCC acted. An FCC-blessed real-world VoIP interconnection trial will not help the Commission clarify the statutory basis for incumbent LECs duty to provide VoIP interconnection. That clarification begins and ends with an interpretation of the statute. 21 19 20 http://www.voip-info.org/wiki/view/free+world+dialup See http://www.att.com/u-verse/explore/home-alarm.jsp (emphasis added): AT&T U- verse Voice service is provided over AT&T's world-class managed network and not the public Internet. 21 Testimony of Commissioner John D. Burke, Chairman, NARUC COMMITTEE ON TELECOMMUNICATIONS, on behalf of the National Association of Regulatory Utility Commissioners (NARUC), before the United States House of Representatives Energy and Commerce Committee Subcommittee on Communications and Technology, hearing on The Evolution of Wired Communications Network, October 23, 2013, at 7-8. 8
As mentioned above, the existing approach of incrementally adding specific regulatory obligations (while avoiding the central question) has reached the end of its usefulness. All major US carriers have concluded that IP technology is the end-state technology for all communication services, including phone services. 22 In order to reach this state, without losing all consumer and competitor protections, the FCC will have to confront the basic question whether its regulatory authority (grounded in telecommunications) will survive the transition to VoIP. 4.0 Conclusion The FCC is currently studying the various issues that arise as wireline networks move to IP technology. Some of these issues impact consumers. For instance, traditional wireline phone service is provided using electrical power on the phone line itself, so that when commercial power fails, the phone network continues to work. IP networks require an intelligent device at the service-delivery point, and these devices require commercial power provided by the customer. While battery back-up is typically provided to ensure consumers retain wireline voice service during relatively short outages, the reality is that extended outages (due to weather or other regional disturbances) will result in a loss of wireline service. Moreover, there are a number of ancillary devices for instance, medical alert devices and alarm services that are premised on the existence of a powered-copper circuit that may not be available after an IP network is deployed, particularly if the transmission medium is replaced with fiber. Of particular importance to competitive providers are issues concerning last-mile access options and interconnection. As in Europe, there is significant competition in the United States dependent upon unbundled access to reach end-users. In the United States, however, there are no packet-based wholesale last mile options (such as Ethernet) upon which a contemporary product line can be developed. Moreover, there is a fundamental dispute as to whether the core interconnection obligations obligations that require incumbents to treat smaller entrants as equals will apply to IP networks so long as the FCC allows the classification question to remain unanswered. When IP was nascent and the future less clear, these disputes seemed remote, with little immediate commercial impact. Today, however, the questions are immediate, even if the answers remain illusive. 22 The term end-state in this context is somewhat exaggerated in that inevitably some other technology will likely replace IP, just as IP is replacing digital TDM services (which had replaced the analog services that came before them). 9