Fiber to the Home -- is too much really enough?
Fred Goldstein, November 2003
...The bottleneck occurs at the LAN-WAN boundary. Telecom bandwidth isn't cheap, especially from incumbent local exchange carriers ...
.
...Few consumers really need more bandwidth than DSL can deliver. ADSL can deliver more upstream bandwidth than the 128 kbps that the Bells usually offer....
...Both FTTH and HFC pull fiber to a place near the subscriber and have an optical-to-electrical converter of some sort. In HFC's case, it's shared among a number of subscribers; FTTH puts one at each home...
...FTTH will only catch on if it has something else to offer, and so far that weapon of mass-media destruction hasn't turned up...
...The FCC's current policies are exactly the opposite of what's needed. Bell-style closed FTTH is the next stage in Bruce Springsteen's old complaint about "57 channels and nothing's on" ...
It's one of the biggest buzzwords in telecom today. Yet it's not even a new one. "Fiber to the home" (FTTH) has been on the lips of telecom industry players since the 1980s. The original CCITT Broadband ISDN project, which begat Asynchronous Transfer Mode networks, started in 1985. Even back then, it was simply assumed that copper local loops were on their way out, glass was coming, and networks needed to be prepared. And the local telephone monpolies already got regulatory concessions that were intended to pay for FTTH deployment. FTTH may well have been derailed, ironically enough, by the short-lived ATM boomet of the early 1990s. The early ATM players focused on the local area network, rather than the public network, and when that business was eclipsed by faster Ethernet-based technologies, the bloom was off the ATM rose. By that time, Digital Subscriber Line technologies were putting new life into the old copper loop. So FTTH went on hiatus. But it's back.
The question, though, is why? Not that there's necessarily anything terribly wrong with FTTH, but what is driving the interest? Is there new demand (a real need for the bandwidth)? Is there new supply (thanks to new technology, is the cost falling rapidly)? Is there a glut of capital looking for the next big thing? Or is something else going on? The classic boom-bust cycle is most severe when supply and investment get ahead of demand. But in this case, there may be little new supply or demand, just an excuse for some political and regulatory manipulations.
A culture of excess
Here in the United States, the consumer economy, indeed much of the popular culture, is largely driven by excess. MTV's old motto perversely captures the mood: "Too much is never enough". I'm not quite sure what that was originally supposed to mean, but the public seems to have taken it to heart. Gluttony is the order of the day at restaurants, many of whom compete by offering huge portions of often-mediocre food. Automobile makers compete by offering more horsepower, while vehicles get bigger and fatter. While some enthusiasts may actually notice the horsepower, most drivers rarely do. Your basic American muscle car has never been well balanced. During the muscle car boom of the 1960s, a 45-horsepower Mini Cooper could still outrace a 250+ hp muscle car ifthe course had enough turns. The automobile industry today again emphasizes hugeness and horsepower. The Pontiac GTO is coming back, and the Mini Cooper has been revived, but now the latter squeezes three times as many horsepower into its tiny body as the original.
In the telecom arena, bandwidth is the equivalent of horsepower. Whether the bandwidth is useful, or even usable, is not the key question; vendors instead emphasis peak burst speed. Of course this is easiest to attain in the LAN. So popular wireless cards for laptops are now supporting the "54G" standard, even though the increased speed works only across very short ranges, and the bulk of applications can't even begin to fill an 11 Mbps 802.11b pipe. Gigabit Ethernet is even supported in some desktop computers. The bottleneck occurs at the LAN-WAN boundary. Telecom bandwidth isn't cheap, especially from incumbent local exchange carriers. So these high speed LANs are all dressed up with no place to go.
Is it a dog-and-cheese problem?
Policy-makers have caught on to this disparity, and have a favorite solution. If only there could be FTTH, then there's be ample bandwidth for all. The economy would improve, unemployment would go down, cancer would be cured, and the Cubbies would win the World Series. Well, maybe they don't go quite that far. But the lack of fiber to the home is taken more seriously than a lack of fiber in that gluttonous fast-food diet. It's the excuse for all sorts of bad ideas.
The fiber shortage reminds me of a particularly silly television commercial from a few decades ago. "Is your dog getting enough cheese?" While that may have sounded like a way to move cheese-flavored kibble, dogs really don't need cheese, and, anyway, I don't have a dog.
Not that FTTH is a bad thing, all other things being equal. Twisted-pair copper is obsolescent, especially the long runs that used to be so common. But not much copper is pulled today anyway. Current telco practice is to only use plain copper for fairly short runs, say up to 12,000 feet from the central office, and only for small customers. Big buildings often get glass already. Longer distances use hybrid fiber-copper, the so-called Carrier Serving Area (CSA) concept, with optical fiber-fed remote terminals feeding relatively short copper distribution subloops. While it's possible to put DSL in the remote terminal, and even feed some video across is, few companies do this; twisted pair is really not a great high-bandwidth medium. But for consumer voice and even realistic levels of consumer data, it's adequate. Few consumers really need more bandwidth than DSL can deliver. ADSL can deliver more upstream bandwidth than the 128 kbps that the Bells usually offer, and other DSL technologies are also quite flexible. So if video is left out of the mix, copper has plenty of life left.
If video joins the mix, FTTH beats twisted pair, but the CATV industry's key technology, Hybrid Fiber-Coax (HFC), is usually more than adequate. HFC can deliver hundreds of broadcast TV channels, and has enough bandwidth for a reasonable amount of video-on-demand. It can support higher data speeds than DSL; the current cable modem standard, MCNS DOCSIS 2.0, can go up to 40 Mbps downsteram and 30 Mbps upstream. That bandwidth is shared by the customers on one or more optical transition nodes, which on new builds, typically serve only a few dozen to a couple of hundred subscribers. It's rarely a bottleneck, and bottlenecks, should they occur, are fairly easy to fix. The PacketCable standard makes cable telephony gear into a dirt cheap commodity, completing the triple play.
I don't understand why telephone companies are so afraid of HFC. A few did try it in the 1990s, but the biggest among them were SNET, Ameritech, PacBell and GTE. The first three were shut down when SBC bought them; Verizon shut down the latter as part of its anschluss with GTE. The closest telco equivalent is VDSL, which can squeeze 50 Mbps onto short copper loops; its short range works best with a fiber-rich HFC-like architecture or "fiber to the curb", in which a small cluster of homes share the DSLAM.
FTTH prices have fallen, but they're not at HFC levels yet. Equipment vendors I've talked to recently suggest that using the emerging Passive Optical Network (PON) standards, it's about a 15% capital expenditure premium over HFC, which could be made up for with lower operating expense. Nice theory, but real-world prices of HFC gear have fallen too, making the disparity quite a bit wider. Both FTTH and HFC pull fiber to a place near the subscriber and have an optical-to-electrical converter of some sort. In HFC's case, it's shared among a number of subscribers; FTTH puts one at each home. Fiber optic drops from a PON are not much more costly than coax; it's the transition to electrical form that makes the difference. Until that component gets cheaper, FTTH will have a hard time competing on price. And since HFC (for cable companies) and DSL (for telcos) meet most of today's demand, FTTH will only catch on if it has something else to offer, and so far that weapon of mass-media destruction hasn't turned up.
The incumbent local telephone companies know this, but they aren't going to admit it. The FCC, in its extremly ILEC-friendly Triennial Review [of network unbundling requirements] Order, has given them a huge gift as an "incentive" to pull more FTTH. If they replace existing copper with glass, CLECs can only get a voice-grade channel out of the new glass. On a new "greenfield" build using FTTH, the Telecom Act's loop unbundling requirement is waived entirely, and CLECs will have no access to the subscribers unless they can pull their own parallel distribution facilities. The FCC's excuse for this? The Telecom Act doesn't speak of "incentives", but of competitive "impairment". What few FTTH lines have been pulled to date have largely been pulled by non-ILECs, therefore "proving" that ILECs don't have an advantage. (Never mind that these FTTH projects have largely been subsidized by local utility districts.) The Bells are in effect being rewareded for their failure to meet their earlier promises! This is obviously generous enough, but some Bells are still complaining about that voice-grade channel. Too much is never enough, it seems.
Content vs. carriage, again
For the public, this is certainly a bad deal. The Bells do not want to have common carrier status for broadband bandwidth on this new glass. Their latest excuse for not building it is that they can only make it worthwhile if they get absolute control over the content as well as carriage. So if they get their way, ILEC FTTH end users will get no choice of ISP, and the Bells, as unregulated information providers, will have the right to choose what web content they can see, what their email will look like, what Internet services they can use, where they can shop on line, etc. These are all "incentives" that a monopolist would like. Residents of these FTTH-served homes will in effect be victims of it, not beneficiaries.
But their dogs will have plenty of cheese.
American consumers should be provided with ample bandwidth for their voice, data and video applications. Within a few years, fiber to the home may become be the economically-right answer for enough of them to make it worth the carriers' while to pull it. It's already a reasonably cost-effective answer for some niche applications, including greenfields, without removing any unbundling requirements. The FCC's current policies are exactly the opposite of what's needed. Bell-style closed FTTH is the next stage in Bruce Springsteen's old complaint about "57 channels and nothing on". Too much bandwidth, without the freedom to use it, is counterproductive. FTTH will only benefit the public and be profitable when it's pulled on a common-carrier basis, or even better, on a neutral outside-plant basis, available on equal terms to any carrier or service provider.
