Optical fiber networks dominate national cable-TV conference

Optical fiber networks dominate national cable-TV conference

GEORGE LAWTON

The National Cable Television Association Show that was held in Dallas last May attracted record crowds from the telephone and cable-TV industries who are interested in building fiber-to-the-curb and hybrid fiber/coaxial-cable networks, respectively, for the next generation of voice, video and data services.

More than 24,350 attendees packed the Dallas Convention Center. Among the show`s highlights was a panel discussion of cable-TV industry leaders, including Rick Meyers, chief executive at Communications Equity Associates, Tampa, FL. He pointed out that the cable industry borrowed more money last year than in the previous two years, partly because of new capital spending on fiber optics and other infrastructures. Private debt for cable systems was $848 million and $1.5 billion in 1992 and 1993, respectively. Last year, cable operators raised $7 billion in new debt, of which $1.5 billion was for capital expenditures.

Another panelist, Richard Green, chief executive at Louisville, CO-based Cablelabs, reported, "All of our companies are focusing on hybrid fiber/coaxial-cable architectures. The regulatory uncertainty has slowed the pace, but it has given us the opportunity to work out the kinks with the architecture. I would say we are still confident that it provides the best economic vehicle for delivering advanced services."

Green also noted that the high end of cable-system development is illustrated by the Full Service Network being constructed in Orlando, FL, by Time Warner Inc. (See Lightwave, April 1995, page 1.) "It is a video-on-demand network that will eventually connect to 4000 homes. However, the project has often been considered too expensive because of its high cost per home."

The network merges cable, telephone, and computer technologies from a team of leading companies, including AT&T Network Cable Systems, Silicon Graphics Inc., Scientific Atlanta Inc., Hitachi America Ltd., Toshiba America Electronic Components Inc. and Hewlett-Packard Co.

In response to these critical comments, Gerald Levin, chief executive at Time Warner, said that the Full Service Network trial was achieving its goals. "Most of all, I think it is significant because we have demonstrated its functionality. You can use a regular cable system that has been upgraded the way most of us have upgraded our systems."

The company has installed video servers, asynchronous transfer mode switches and computer workstations disguised as set-top boxes in people`s homes. Levin noted that the software has been the toughest integration task, and that the trial has required more lines of code than were written by NASA in the first trip to the moon.

However, the key limitation now is how inexpensively these components can be integrated into other systems. "I think we found that the silicon math is the driving force of our industry," Levin said. Silicon math refers to the rise in performance and the drop in cost of silicon components, which is accelerated by the volume deployment of set-top boxes. The next generation of video games will use the same silicon as that placed into set-top boxes, which will help increase the volume of silicon produced while reducing its price.

Increased growth of the telecommunications industry has resulted in lower costs for other equipment. Time Warner, for example, paid $3 million for the ATM switch used in the Orlando trial. But 18 months later, a similar switch costs only $100,000.

Leven notes, "I believe we will have a $300 set-top box at the appropriate moment, which is defined when we know what the consumer wants. The box will be able to summon everything, including movies or news."

The National Cable Television Association gathering also showcased new products. Among the devices was an Ortel Corp. linear return-path laser for cable systems that intend to offer such bidirectional services as telephony and high-speed Internet access. Because cable operators are used to working with linear fiber optics for video transmission, a linear return-path laser fits into their architecture, says Larry Stark, Ortel`s vice president and business manager of broadband communications. In fact, most of the hybrid fiber/coaxial-cable network architectures being proposed by such vendors as ADC Telecommunications Inc. and Antec Inc. are using linear lasers for the return path.

The functional demand on return-path lasers is not as great as it is on the forward-path laser because they are not typically driven over such a wide range. A typical downstream laser transmitter has to accurately transmit from 50 to 750 megahert¥or even to 1 gigahertz. The radio-frequency return on cable systems is typically in the 5- to 50-MH¥range. In some cases, a low-cost digital Fabry-Perot digital could do the job, according to Stark.

Fiber goes deep

In addition to its work on new interactive technologies, Time Warner has been studying the financial and reliability impacts of different fiber architectures. In mid-1994, the company launched a test bed in Monroeville PA., a suburb of Pittsburgh, to study a fiber-intensive architecture known as Fiber Deep. The system comprises 330 miles of plant with an average density of 145 homes per mile. Two hundred miles of the system are being upgraded using the Fiber Deep design.

Don Gall, a senior project engineer at Time Warner in Denver, described the company`s fiber-deep architecture. Time Warner is currently using a coaxial-cable bus architecture known as Fiber Rich. This design calls for a maximum cascade of two trunk amplifiers, one bridger and three line extenders. This design has been cost-effective and enables the company to upgrade a typical 300-MH¥system to 750 MHz, at an average cost per mile of $12,000 to $14,000.

A fiber-deep architecture pushes fiber all the way to the bridger amplifier, resulting in a lower number of RF amplifiers for the network. These amplifiers are often regarded as the weak links in a cable system, because they tend to fail more often than fiber-optic components. For comparison, the Fiber Rich architecture had 6.4 amplifiers per mile compared to 5.21 amplifiers per mile for Fiber Deep.

Gall noted, "The reduced cascade allows for higher output levels. These higher levels permitted fewer amplifiers per mile than required in the Fiber Rich architecture. Besides being reliable, the design also lends itself to the implementation of the hybrid fiber/coaxial-cable network powering of telephony. We feel that the concept should be cost-effective for the 70 to 80 homes per mile densities in plants where the trunk is 90% or more aerial." q

George Lawton writes from Brisbane, CA.

More in Transport