Transmission pipes: loops and transport

Loops are so-called “last mile” facilities: the wires or cables a telecommu-nications company uses to connect its customers to the nearest switch and thence to the rest of the world. Strung through the air or laid underground (sometimes in tubes called “conduits”), loops constitute by far the costliest portion of most telecommunications networks and are thus the most diffi-cult facilities for an upstart company to duplicate. For that reason, they are often described as “bottleneck” facilities, with major regulatory conse-quences.

The basics of wireline transmission

The most traditional form of the loop—a typical telephone line—sists of a twisted pair of copper wires used to establish an electrical

con-*When reviewing the following discussion, readers may wish to glance forward to Fig. 1, which appears on p. 41 at the conclusion of our description of the tradition-al telephone network.

nection with the telephone company’s switch. When a customer lifts the receiver off the hook to place a call, the switch sends her a dial tone to con-firm that the circuit has been established and is available to carry her call.

When someone else calls her, the switch sends another electrical current down the line, this time to trigger the ringing of her telephone. The electri-cal currents come from giant batteries at the telephone company’s switch-ing station, known as a central office—usually identifiable as a bland-looking building with few windows and the telephone company’s logo outside. For public safety reasons, the electricity that powers ordinary telephone service is independent of the electricity that provides power in the rest of your home. Thus, when the electric grid fails, you can still use your phone unless, of course, it is a cordless one whose base station requires ordinary AC power.

Suppose that you place a telephone call to someone whose loop is con-nected not to your local switch, but to a neighboring one. To get from your switch to hers, your voice signals must travel along a high capacity trans-port link. Most modern transtrans-port facilities use optical fibertechnology; to some extent, as explained below, telephone companies have begun using that technology in loop facilities as well. In effect, a fiber-optic cable is an extremely thin glass tube that transmits light over long distances through various forms of internal reflection. Laser-originated light waves, carrying signals, bounce from one end to the other. (One common, though quite rough, analogy compares fiber optic technology with shining a flashlight down the interior of long tube with a mirrored surface on the inside;

although the tube may bend and twist, the light shines out of the other end.) Attached to each fiber strand are expensive electronic devices that aggregate all of the signals from different customers’ individual lines onto the same physical strand of optical fiber. This is called multiplexing.

Copper wires often carry multiplexed signals too, but not with the phe-nomenal capacity of fiber.

In the traditional wireline telephone world, the most common form of multiplexing—known as “TDM” for time division multiplexing —“sam-ples” the signal for a given call many times a second and transmits those samples along with the corresponding samples taken of other calls. A “sam-ple” is a kind of digital snapshot of the signals in a call at any given moment, much as a frame in a movie is a snapshot of the action in progress. Each call is preassigned time slots in the multi-call transmission;

at the other end, this aggregated signal is “de-multiplexed” back into

indi-Introduction to Wireline Telecommunications 35

vidual signals. By means of this multiplexing technology, a single strand of fiber thinner than a human hair can carry thousands of simultaneous voice conversations. This aspect of telecommunications technology often comes as a surprise to people outside the industry: the calls you place to your friends across the country typically coexist on the same fiber strand with many other calls taking place between people you don’t know.

The use of optical fiber is routine on any transport route where extremely high numbers of calls need to be transmitted at once, such as between cities or between most central offices. The use of fiber to connect central offices to customers—i.e., in loops—is less predictable. Sometimes, optical fiber is used only in the aggregated multi-loop feedercables closest to the central office, and copper wires remain the medium of choice for the more diffuse distribution portion of the loops closest to the customers.

(Think of feeder cables as the main branches of a tree and of distribution cables as the smaller branches and twigs.) But telephone companies some-times use optical fiber all the way from the central office to the locations of certain customers with very high call and data volumes, such as large businesses.

Whether it makes economic sense to use optical fiber rather than cop-per wires in new loop facilities is often a complex question. Fiber provides more capacity and lower long term maintenance costs, but also consider-ably higher up-front costs in the form of the necessary electronic equip-ment. The choice between copper and fiber thus depends on, among other things, the length of the loop and the “line density” of a particular area:

i.e., the number of homes and businesses in close proximity. Fiber-optic technology is most prevalent in the major cities, where fiber “rings” pass beneath the streets in downtown business districts to collect the enormous call volumes coming from large office buildings. On a smaller scale, fiber also has become increasingly popular in densely populated residential areas, both because its long term maintenance costs are lower and because it enables telephone companies to provide broadband Internet services to more customers, and at faster speeds, than is possible using purely copper loops.

The last point warrants some elaboration. Fiber has exceptionally high bandwidth—i.e., data carrying capacity—that does not vary significantly with the distance between the telephone company’s central office and a customer’s home. One industry study found that an all-fiber (“fiber-to-the-home”) loop would enable residential customers to download a high

qual-ity copy of the movie “Braveheart” onto their computer hard-drives in less than half a minute (if only there were commercially available computers capable of keeping up with the task).¹The bandwidth of copper wires is much more limited and varies inversely and dramatically with length. The major “broadband” Internet access service now offered over copper loops is known asdigital subscriber line(DSL) service, which we will discuss in chapters 4 and 5. The speediest, most expensive variants of DSL—which still offer only a fraction of fiber’s bandwidth—are available only to cus-tomers living in very close proximity to the telephone company’s central office or other specialized data-handling facility. And even the slower, more standard versions of DSL are generally unavailable to customers more than 18,000 feet away.

The fiber glut

Not too long ago, media coverage of the telecommunications industry focused on the “fiber glut” that sent high-flying long distance companies such as WorldCom (now MCI) and Global Crossing into bankruptcy.

Along long distance transport routes, this glut is real,²and is the result of two basic factors. First, in the 1990s, many companies racked up huge debts laying redundant fiber-optic cables over the same city-to-city routes on the mistaken—and, in retrospect, wildly unrealistic—assumption that demand would keep pace. That assumption was memorably encapsulated in a turn-of-the-millennium Qwest commercial, in which a weary traveler shows up at a hotel in the middle of nowhere, asks the desk attendant what movies are for rent, and learns that, through the miracle of fiber optics, he can watch (on demand) any movie ever made in any language.

In reality, such exponential growth in the demand for long distance telecommunications capacity would take many years to materialize, in part because too little fiber has been deployed in the last mile to generate demand from individual customers for bandwidth-consuming products, and in part because content providers fear that making such products available on the Internet will lead to widespread illicit copying.³ As a result, people have continued renting movies in hard copy or watching them on cable (or satellite) television—and are likely to keep doing so in the near term. In the not-so-distant future, the broadband vision depicted in the Qwest commercial may well become a reality, at least in some areas.

Making that vision a reality, however, will require overcoming obstacles

Introduction to Wireline Telecommunications 37

both (i) to the deployment of high speed connections to a critical mass of customers and (ii) to the wide dissemination of digital content that will encourage customers to order those connections.

The second primary reason for the fiber capacity glut is technological.

In recent years, advances in “wavelength division multiplexing” have dra-matically and unexpectedly increased the signal-carrying capacity of opti-cal fiber, including some fiber that had been in the ground for years.

Among other innovations, engineers have greatly expanded the bandwidth of a given fiber strand by finding new ways to send signals within that strand simultaneously over multiple wavelengths of light—colloquially known as “colors,” even though the signals at issue lie outside the visible spectrum. The result of such technological advances, combined with the manic overinvestment of the late 1990s, led to an extreme surplus of dark fiber, so named because no electronics have been placed at either end to

“light” the fiber up with laser-guided signals. True to the laws of supply and demand, this overcapacity slashed the rates that debt-saddled telecom-munications carriers could charge for providing transmission services along the major routes.⁴

Significantly, the routes subject to this overinvestment were those with the highest volumes of conventional voice calls and Internet traffic: the long distance transport routes between major population centers. Again, there was little fiber overinvestment withinany community, except for the most densely populated, and there is no fiber glut—and sometimes no fiber at all—in the last mile loopfacilities to most homes and small businesses.

Those facilities retain some “bottleneck” characteristics in many areas, at least if one does not count the alternative last mile facilities of cable or wireless companies. Policymakers have thus tended to agree that competi-tion in the provision of local wireline telephone services would develop quite slowly for all except the highest volume customers unless new entrants enjoy a regulatory entitlement to rent at least some of the loop facilities of established telephone companies. By contrast, because a num-ber of national providers own long distance transport networks, a thriving wholesale market for leased capacity on those networks arose without heavy government intervention. As discussed in chapter 3, many of today’s regulatory battles about leasing rights are waged in the territory between these two extremes: in the market for “local transport,” including the links connecting two of a telephone company’s central offices.

Regulatory distinctions among transmission pipes

People in the telecommunications industry use broad-brush terms like

“local” and “long distance,” and “loops” and “transport,” as useful short-hands to describe the different ways in which transmission pipes are used in telecommunications networks. But it is important to keep these distinc-tions in perspective. As we shall see, regulators attach great significance to the difference between “local” and “long distance” services. Until quite recently, they fenced the largest “local” wireline carriers out of the “long distance” market altogether, and they still draw bright geographical lines to identify when a call qualifies as “long distance.”⁵From an engineer’s per-spective, however, there is no clear demarcation point between “local” and

“long distance” transport: pipes run the spectrum from long to short and from higher capacity to lower capacity. This is notto say that regulators have wholly contrived the distinction between local and long distance serv-ices on the basis of arbitrary criteria. Distance remains relevant from a business and technological perspective insofar as it corresponds to com-mercially significant phenomena like traffic volumes and their associated economies of scale. But once a telecommunications network is up and run-ning, a carrier incurs little, if any, extra cost to send a call a thousand miles to its destination rather than ten miles.

The rigid long distance-local distinction that characterized the market in the last two decades of the twentieth century is increasingly giving way to these underlying economic and technological realities. As local telephone companies enter the long distance market, and as long distance companies enter the local telephone market, the distinction between these two markets has inevitably blurred. As in wireless markets, wireline consumers are increasingly able to purchase pricing plans with large buckets of inter-changeable “long distance” and “local” minutes. And the providers of

“voice over Internet protocol” (VoIP) almost always offer consumers the same option.

There is also no firm engineering distinction between a “transport”

facility and a “loop.” Remember that both are simply transmission pipes.

By definitional tradition, a “loop” connects a switch and a customer loca-tion rather than, as with a transport link, a switch and a switch. But sup-pose that the customer connects its end of the loop to a switch of its own.

For example, the customer may be a large business with its own private

Introduction to Wireline Telecommunications 39

internal switch, known as a private branch exchange(PBX); or a compet-ing telecommunications carrier with its own switch; or an Internet service provider with a “modem bank” connected to a router that in turn leads into the Internet. In each of these cases, the distinction between “loops”

and “transport” breaks down somewhat. It is useful to keep this point in the back of your mind as you think about telecommunications issues, even though regulators normally designate the facilities as one or the other.

One final point before we move on to switching technology: the loop facilities described in this chapter are the main ones used by “wireline”

telecommunications carriers, a somewhat arbitrary category consisting of the companies that built wired networks for the primary original purpose of providing point-to-point voice and data services. But cable companies now lead the residential market in the provision of high speed point-to-point dataservices, and they have begun entering the market for point-to-point voice services as well. As discussed in chapter 4, their “loops” are the same cables—usually a combination of optical fiber and, in the portion nearest the customer, coaxial cable—used to bring television signals into American homes.

Conventional wisdom holds that, to prevail in the market for fixed (non-mobile) residential services over the long term, a communications provider must find a way to offer consumers the so-called “triple play” of voice, data, and video services. The battle for this market—which ultimate-ly pits wireline telephone companies against cable companies—will likeultimate-ly be won by the providers that find ways to cover the exorbitant costs of pushing more fiber-optic cable into more residential neighborhoods closer to the ultimate consumers. We will defer a full discussion of that cross-plat-form competition, and its highly controversial regulatory dimensions, until chapters 4-6, below. And, in chapters 7 and 8, we will address another crit-ical alternative means for bridging the last mile to the customer: wireless connections between a customer’s mobile telephone and the nearest tower on her cellular network.

B. Switches

A network is defined by its switches (or routers, as they are typically called in the Internet world). Imagine trying to connect every home or business in the United States to every other home or business without the use of a

switch. The number of required lines, and thus the cost, would be astro-nomical. In fact, if we estimated the number of wireline telephones in the United States as roughly equivalent to the number of Americans, the tan-gled mess of lines criss-crossing the country to connect each telephone to every other would amount to more than 40 quadrillionlines.⁶

Switches are built to solve this problem in the most economical way.

They direct a voice or data call from one transmission pipe (a loop or transport link) to another en route to the call’s destination. Although the distinction can blur at the margins, there are two basic kinds of switches—

circuit switchesand packet switches—that are used, respectively, in conven-tional voice networks and more advanced data networks, including the Internet. As mentioned, the physical infrastructure of wireline telephone networks overlaps significantly with that of the Internet. The major excep-tion to this rule lies in switching technology, for reasons we discuss below.

Circuit switches

Circuit switches include the early hand-operated switchboard and its modern-day functional equivalents in virtually every conventional tele-phone network. A circuit switch sets up a dedicated transmission path from the calling party to the recipient for the duration of a call. At any point dur-ing the call, a particular increment of capacity is reserved for that call on the loop, switch, and transport pipe, even if no one is talking and no infor-mation is being sent.⁷To save money, a telecommunications carrier does not build enough capacity on its switches and inter-switch transport links to carry calls from all customers at once. Instead, like a bank, it keeps just enough in reserve to cover the greatest reasonably expected demand. The size and cost of switches and transport links are thus determined by the expected capacity needs of the network at peak calling hours. This is one reason why many callers in the nation’s capital received “all circuits busy”

signals when calling home on September 11, 2001: telephone engineers had not built in enough network capacity to serve this unexpectedly high call volume.

Modern circuit switches are essentially very large computers that, in addition to establishing circuits for given calls, perform a variety of other

“intelligent” functions, including call forwarding, caller identification, and call waiting—known collectively as vertical switching features—as well as billing. A modern circuit-switched network is usually shadowed by a

par-Introduction to Wireline Telecommunications 41

allel, packet-switched signaling network, which tells the circuit switches how to route particular calls to avoid network congestion and how to implement specific customer requests, such as where 800 number calls should be directed and how calling card calls should be handled. The

“brains” of a circuit-switched network are said to reside in the switch and the parallel signaling network, not at the “edge” of the network in an end user’s computer, and they are centrally owned and controlled by the

“brains” of a circuit-switched network are said to reside in the switch and the parallel signaling network, not at the “edge” of the network in an end user’s computer, and they are centrally owned and controlled by the