Goal is broadband at less than $1,000 per sub

Fred Dawson, formerly editor and publisher of the highly regarded Cable-Telco Report, has joined Digital Media as contributing editor.

Local telephone exchange carriers are seizing on new technical approaches to delivering video and other broadband services that could lead to a much faster rate of network expansion than once seemed possible.

Buoyed by the prospects for new business growth under less restrictive regulations (see Vol. 3, No. 4, p. 3), the regional Bell operating companies and major independent telephone service providers are rethinking traditional fiber-based designs in an effort to drive broadband network costs below $1,000 per household, which is less than half the costs once contemplated for such expansion.

QUESTIONING THE ALL-FIBER ASSUMPTION

For years local exchange carriers (LECs) assumed that an all-fiber approach was the route to the broadband future. Fiber optics, with its virtually unlimited bandwidth, would eventually replace the traditional twisted pair copper wires, providing a transmission medium that accommodated the standard plain old telephone services (POTS) and a full range of new services, from video conferencing and multimedia to high-definition television.

Under this scenario, the high-capacity fiber would map to the replaced copper network topology all the way to the home, so long as regulators could be convinced that sheer aging of existing wiring (referred to as “plant”) or new market demands for higher capacity services or some combination of both justified the cost of adding fiber without creating upward pressure on standard telephone service rates.

Costly conversions. But driving fiber all the way to the customer premises requires an optoelectronic conversion at each household to accommodate signal transport over copper and coaxial cable links within the home. About five years ago it became evident that costs of such conversions at each household would unreasonably delay the evolution to higher capacity. This led to industry-wide acceptance of the idea of an optoelectronic conversion at an optical network unit (ONU) serving a cluster of eight or so premises from a curbside cabinet, via twisted pair copper drop lines for voice circuits and coaxial cable for video and other broadband services.

Now, even this fiber-to-the-curb (FTTC) scenario is viewed as too costly by a growing number of telcos, who fear that regulators will balk at allowing them to incur such costs without a better understanding of how new services rather than higher voice service rates will defray the expense. While FTTC costs are coming down, other approaches have been found that appear more likely to satisfy regulators’ insistence that network expansion costs pay for themselves.

Following cable’s lead. The leading alternatives to FTTC include fiber/coaxial configurations similar to cable television industry network designs, short-haul wireless access links and high-capacity copper wiring. Increasingly, telcos are not so much looking for a single design solution as they are seeking to match particular design attributes to specific network conditions, including variations in the age of the plant as well as the locale, whether it’s urban, suburban or rural.

“Our goal is to identify the network design best suited to a particular serving area, which could mean different approaches are taken from territory to territory,” says Larry Plumb, spokesperson for Bell Atlantic. “The key thing is that all the last-mile elements must be compatible with the operating core.” By this he means that alternative means of distribution to the customer must connect economically with the existing array of switches and the standard operating protocols of the telephone industry.

A key advantage to the new flexibility in network design is that the placement of broadband capacity does not necessarily have to depend on obsolescence of existing copper plant. The way the game is played today, if an area requires new telephone lines, a large share of the costs of installing the fiber that will support broadband services can be shifted over to the narrowband side, which helps overcome cost-conscious regulators’ resistance to capacity expansion.

But because the age and condition of last-mile telephone wires varies widely, even within small areas it’s hard to find a rehabilitation scenario large enough to support the scale of construction that’s required in a conversion to fiber.

ALTERNATIVE NETWORK DESIGNS FOR “SERVING AREAS”

The leading network design alternative in the telcos’ search for a way around the narrow/broadband cost linkage is the fiber/coaxial “fiber-to-the-serving-area” (FSA) topology being championed by the cable television industry, where fiber extends to coaxial serving areas of 500 to 2,000 homes each. Such a network can be overlaid on an existing telephone network, connecting at key switching points and sharing trenches and conduit, without matching the layout of the twisted pair copper network line by line.

According to John Czak, executive director of network systems and planning at US West Communications, the FSA “star/bus” approach costs less than FTTC because the coaxial network beyond the fiber termination point or node follows a “bus” distribution concept, where one line goes down a street, serving all homes via short connecting “drop” extensions.

This contrasts with FTTC, where individual fiber links extend to each eight-home cluster all the way from the switch, which can be three miles or more away. Thus, there are far fewer aggregate miles of distribution cable in the fiber/coaxial scheme than in the FTTC approach.

“We’re getting a message from our customers that’s loud and clear,” notes Larry Burton, director of advanced technology and products at Reliance Comm/Tec, a leading supplier of last-mile distribution equipment. “If the consumer is getting what he wants over a 30-channel cable system, that’s the price to beat if you’re a telephone company.”

REGIONAL BELL COMPANIES MOVE TOWARD FSA

US West began the trend toward telco FSA late last year, saying it would deploy FSA broadband at a pace of 500,000 homes per year, following a year’s worth of testing in 1994. The first iteration of the system, will serve 10,000 households in Omaha, NB, starting early next year.

Groundswell of interest. Rochester Telephone of New York and Southern New England Telephone have also chosen FSA topology for initial tests of video services, although they have made no commitments to the design beyond the initial trial stages. But vendor sources say there has been a groundswell of interest among most of the major local exchange carriers during the past few weeks, with the likelihood that several others will at least go to the trial phase using FSA architecture.

Sources also report that Ameritech has distributed a formal solicitation for information on FSA designs, laying groundwork for a possible request for proposals (RFP) from vendors later this year. Nynex is also taking a hard look at widescale deployment of fiber to the service area, according to the same sources. Neither company would comment other than to say they are looking at a variety of networking options.

Cable’s logic hard to deny. “It’s getting hard to deny the argument for the cable FSN (full service network),” says a strategist at a telco that hasn’t declared its design preference. Requesting anonymity, he adds, “Broadband is where the new revenues are, so why hogtie expansion to rebuild schedules?”

Czak of US West says another big advantage for FSA is that it supports low-cost “broadcast” delivery of analog programming networks, which comprise the lion’s share of video services that would be available over a video dialtone network anytime in the near future. Where FTTC broadband is all-digital and all-switched right out of the chute, FSA supports gradual evolution to such capabilities as market conditions require, Czak says.

Coax can supply power, too. Ron Foster, vice president for telecommunications at Scientific Atlanta, a cable equipment supplier that some believe will provide FSA gear to US West in Omaha, also notes a key advantage for FSA can be found in the ability of the coaxial plant to supply electrical power to the customer’s telephone, which fiber can’t do. “The telephone companies have gone down a lot of rabbit holes trying to find a cost-effective way to power CPE (customer premises equipment; i.e., telephones),” Foster says, “but nothing has been found that matches the ability of coax to deliver service and power together.”

“The important thing to recognize,” he adds, “is that FSA can be evolved to do all the things commonly associated with telephone broadband. This isn’t about making compromises with the future.”

WIRELESS ACCESS GAINING POPULARITY IN TELCO CIRCLES

Another low-cost alternative to FTTC that is beginning to gather steam in telco circles is wireless access, where radio transceivers (transmitter/receivers) with a serving radius of anywhere from a few hundred feet to half a mile would communicate with transceivers mounted on exterior walls of customers’ homes and offices. Nynex has been testing narrowband wireless access in urban and suburban settings, with calculated installation and maintenance savings of between 10 and 20 percent as compared to copper access lines and considerably more as compared to FTTC.

But while radio connections might reduce the cost of replacing wires for traditional phone service, the larger advantage to radio lies in the ease with which such connections can be upgraded to full broadband capacity. Instead of new wire installations, the process would merely require modular add-ons at the transceiver sites, notes Paul Donovan, staff director at the Nynex Telesector Resources Group.

PCS opens a conduit. The Federal Communication Commission’s recent provision for telco use of a portion of the newly allocated Personal Communications Services (PCS) spectrum in the 2 GHz region opens a conduit for narrowband wireless access. In this scenario, customer premises equipment might be connected by wire, rather than by radio, to the outside wall-mounted transceivers linked directly to a neighborhood microcell. But the spectrum for moving to broadband wireless access would be found in another region of the microwave spectrum, at or beyond 28 GHz. (For more on PCS, see p. 21)

These very high frequency areas have suddenly come open to mass communications, thanks to the pioneering work of CellularVision, which recently gained Bell Atlantic as a partner in construction of an all-wireless multichannel TV service in New York (see Vol. 3, No. 4, p. 3). The availability of these high frequencies, largely unused in terrestrial communications, has driven the wireless access concept to the front burner at US West, Nynex and Bell Atlantic.

Shorter distance, lower cost. Bernard Bossard, a partner in CellularVision, says that the shorter propagation distances of these frequencies can translate into lower costs for transmitters, which with advances in semiconductor technology could conceivably dip down into the $20,000 range from about $125,000 today.

Such a transmitter is capable of delivering all forms of communications, including voice, data and video in one-way or interactive mode, given a relatively clean line of sight between transceivers.

US West has been testing a prototype broadband wireless access system operating at 28 GHz, according to company sources. Results so far have put wireless access on the short list of possible options as the RBOC moves to widescale roll out of full-service networks in the next year or so.

CONNECTIONS AND COST-PER-HOUSEHOLD KEY ISSUES

A key issue concerns positioning the transmitter to ensure dependable connectivity to all customers in an area. At average suburban densities, a one-mile wide cell, covering a little less than one square mile of territory, would serve about 225 households, which would result in very low costs for the last mile of access — something on the order of $300 per household, counting customer premises gear. A lower power cell transceiver, with radius of a few hundred feet, would cost less than the one-mile cell system, but, with fewer homes served, would produce a somewhat higher cost per household.

Such costs, of course, don’t factor in the costs of bringing fiber to the transceiver site. The larger the transceiver serving area, the less fiber required, so the range of cost possibilities for wireless access is fairly broad. But it appears to fall under the wireline alternatives by a considerable margin.

ADSL getting better. Rounding out the search for cheaper routes to broadband is the widely reported technology known as ADSL (asymmetrical digital subscriber line), which employs advanced modulation and compression techniques to expand the capacity of existing copper wiring. The name refers to the ability to deliver high-bandwidth data such as video into the home, while preserving a much smaller amount of bandwidth for the return — i.e., interactive — path. Here again, Bell Atlantic is among the leading experimenters, with a video dialtone trial using ADSL now underway in the suburbs of Washington, DC.

Bell Atlantic believes ADSL offers a way to bring movies-on-demand, multimedia and other advanced services to the market immediately on a customer-by-customer basis, without waiting for fiber loop deployment. “This is a technology you pay for as the customer requests it, which is a very cost-effective approach to introducing new services,” says Arthur Bushkin, president of Bell Atlantic Information Services.

ADSL, which has drawn sharp criticism from Southwestern Bell and a lukewarm response among several other telcos, keeps getting better. Bell Atlantic is the only company committed to a market trial, but its efforts are viewed as pivotal to go/no-go decisions at a number of other companies.

Massive multitasking with DMT. Earlier this year the Bell companies’ task force on high-speed copper lines endorsed an ADSL variation known as discrete multitone (DMT), which its developers say can deliver four 1.5 Mbps one-way (or two 3 mbps) video signals along with two ISDN B channels, a data channel and an analog voice channel simultaneously over a single twisted pair copper line at distances up to 12,000 feet. In late August, the standards group went even further, endorsing a 576 Kbps channel for high-quality videophone connections along with a 64 Kbps control channel on top of the video channels and analog voice.

These loop distances are sufficient to cover the lion’s share of dense urban markets and represent over 50 percent of all local distribution plant. The technology supports lower levels of capacity at longer distances, with a combination of at least one 1.5 Mbps video and voice and control channels at distances up to 18,000 feet, generally the maximum in telco serving areas.

Not so cheap. Naysayers contend the service isn’t flexible enough and is likely to cost more to install than optimists realize, making it harder to shift capital to fiber deployment when broadband becomes a justifiable expenditure. While the four 1.5 or two 3 Mbps video channels can serve separate TVs with different programming, these rates are below the threshold for NTSC television as envisioned in the emerging MPEG-2 standard. The technology can also support one 6 Mbps channel, which would cover NTSC, but one channel per household is viewed by many strategists as a non-starter.

A final draft on ADSL, which will permit full scale manufacturing to proceed, is slated for the second quarter of 1994. A number of telecommunications manufacturers are gearing up to compete for the anticipated business, including the three largest — Alcatel, AT &T and Northern Telecom.

$250-per is hard to beat. What that cost will be on a per-line basis, Bell Atlantic’s Bushkin won’t say. The goal is to get DMT onto a single chip, which means every household box would have a DMT chip and a decompression chip or, possibly, a large-scale integration of both, as well as sufficient computer power to support customer navigation through thousands of viewing and data source options.

The cost increment at issue vis à vis some other digital networking option is the DMT portion, which, given cost trends, might turn out to be in the neighborhood of $500, counting hardware at both ends of the line. Assuming a 50 percent customer take rate, this would work out to $250 per household. In an all-digital video dialtone mode this might be hard to beat, even for wireless access.

Bushkin and other ADSL advocates stress that the technology is not meant as a replacement for fiber in the loop but rather as a step along the way. As customer demand for initial video services offered over copper loop hits stride, he says, the telco will have the revenue justification for extending fiber and broadening the service options.

Fred Dawson