Broadband connections that provide access to the public network are particularly critical as well as challenging for the home and small-business customer. (Large-business customers can already be served with fiber directly to their buildings.) Homes and small businesses will be the major growth markets for interactive data and video services, and access network products are expected to dominate optical communications equipment sales throughout the next 15 years. By 1998, equipment for fiber-optic broadband systems is forecast to constitute two-thirds of all optical communications equipment sales, expected to total $12 billion; and by 2003, fiber-optic broadband equipment is forecast to constitute three-fourths of $30 billion in total optical communications sales (OIDA 1994). (The corresponding worldwide markets for all optoelectronics, including displays and storage, are $140 billion and $230 billion, respectively, in those years.) In addition, fiber-optic broadband systems also drive development and sales of new home electronics equipment and services.
The ideal access solution would be to use fiber all the way to the home, but fiber-to-the-home products tested and even developed in the past proved too costly. Manufacturers turned to developing other, short-term solutions that reduce costs by sharing components, cables, and installation among several to many customers. These are fiber-to-the-curb (FTTC) and hybrid fiber/coax (HFC). Both bring fiber near groups of customers, carry out the necessary optical/electrical conversions and other functions at a shared equipment location, then complete the connections to customers using conventional metallic cables and conventional network configurations.
Cost is the driving factor - not technology - and is an extremely complex issue, too complex to discuss in any detail here. The bottom line is that HFC is the lowest-cost solution for cable television operators because they have an installed base of coaxial cable, some of which can be reused. For telephone or other operators who must build new networks, cost can be very similar, however. It is becoming apparent that FTTH has become much less expensive recently, approaching FTTC and HFC within only a few hundred dollars. Considering the advantages of FTTH and cost savings that result, many people believe FTTH will be an important factor again very shortly.
Several manufacturers, including AT&T, are currently supplying FTTH equipment for advanced trials in Japan. The goal is to deploy FTTH equipment and services to (or near to) 20% of Japanese homes by the year 2000 and to all Japanese homes by 2010. The Japanese government is to be instrumental in financing approximately $400 billion over the next 15 years to bring broadband access to all Japanese homes (Wall Street Journal 5 December 1994, A-11). There is already some doubt that the timetable will be achieved. The cost of rebuilding after the Kobe earthquake may impact funding of this broadband access infrastructure - effects on the FTTH timetable are not yet known. Also, Japan's Ministry of International Trade and Industry (MITI) has argued that the private sector should pay for the infrastructure construction (Wall Street Journal 15 August 1994, A-1).
In the United States and Europe, mass deployments of systems for broadband access are also beginning, but almost entirely with FTTC and HFC schemes (United States, most of Germany, and other European and Pacific Rim countries). There is no worldwide standard for delivering broadband communications to the home. One of the OPAL '94 deployments in Germany, and trials in England, Denmark, and Germany, use FTTH. Also, early installations in the United States were all FTTH. The U.S. emphasis shifted to FTTC and narrowband delivery about 1989, but then shifted back to video following the Federal Communications Commission's Video Dialtone rulings in late 1991. To address the U.S. market, Fujitsu recently announced a Sonet-based (that is, the SDH standard) FTTC system combined with one-way HFC for broadcast video and powering of the FTTC. For systems using two-way HFC for all services, such as committed to by, Pacific Bell and Southern New England Telephone, NEC recently introduced a telephony-on-coax product. (There are currently fifteen other manufacturers with similar products.) HFC systems principally rely on analog (subcarrier) video transmission over fiber; at least six Japanese companies (Fujitsu, Matsushita, Mitsubishi, NEC, Sumitomo, and Toshiba) make the required linear lasers.
Fujikura is one of several Japanese companies that are now expanding into transmission systems. Fujikura previously manufactured cable, components, and test equipment, rather than the systems themselves. Specifically, that company's emphasis is on the access-network product, and the panel saw Fujikura's NTT-compliant FTTH product in the course of its visit. The panel saw a nearly identical system at Furukawa.
Fujikura's network access product combines two-way ISDN delivery at 1310 nm with video delivery at 1550 nm, namely 11 channels of analog AM-VSB on carriers between 90 and 400 MHz, and 36 channels of FM video plus 60 channels of QAM digital video on carriers between 500 and 2400 MHz. These allocations are compatible with blocks of DBS intermediate frequencies so that the set-top converter can leverage off of available technology. The optical network unit at the home is approximately 4 in. x 8 in. x 2 in., dissipates 7 W, and is available in either one- or four-output versions. By using several stages of amplifiers and optical splitting, Fujikura can use one analog DFB laser to provide service to approximately 10,000 subscribers. Its prototype uses dispersion-compensated fiber, but Fujikura representatives indicated to the panel that they feel another approach to dispersion compensation is needed, since the present method requires inserting 25 to 33% of the span length in high-dispersion fiber.
In the development of FTTH systems, Japan is ahead of activities in the United States, particularly in the areas of small optical network units, subcarrier video at 1550 nm combined with switched services at 1310 nm, and optical amplification of the video signal for subsequent large passive splitting. One exception is AT&T, which, as mentioned earlier, is one of the manufacturers delivering FTTH equipment to NTT. According to the U.S. trade press, AT&T has made significant progress in reducing costs for the Japanese product (Fiber Optics News 10 October 1994 and 6 December 1994). No other vendor in the United States is currently working on FTTH.
The U.S. focus is either on fiber-to-the-curb or two-way hybrid fiber/coax, as mentioned earlier, and in both of these areas, Japan is behind. Importantly, a growing number of people in the U.S. telephone industry are reexamining when using FTTH might make economic sense; despite higher initial costs, FTTH technologies have the advantages of lower maintenance and powering costs and greater upstream bandwidth potential than other alternatives. If in the United States a shift back to FTTH occurs in the next few years, which is very possible since the price differential between FTTH and other alternatives has narrowed substantially, most U.S. manufacturers would find themselves in a catch-up mode.
At the component level, low-cost loop (or "access") lasers and modules, already a focus of intensive development by Fujitsu, Furukawa, Hitachi, Mitsubishi, NEC, and Toshiba, are undergoing further improvements and cost reductions. NTT has set a target price of 5000 yen (about $50) for a transmitter/receiver pair, with the transmitter incorporating the packaged laser module. (This is consistent with a $30 target price for the laser module set by Bellcore in the mid-1980s.)
Loop lasers operate with low threshold currents over the wide outside-plant temperature range of -40 degrees C to +85 degrees C without thermoelectric (Peltier) cooling. Either the packaging or the laser chip itself is designed to relax laser-to-fiber coupling tolerances so that output remains stable over these temperatures. The laser output exclusive of coupling is maintained constant using negative-feedback control of the laser bias, based on sampling the chip's light output using a photodetector package with the laser. In the case of new low-threshold-current MQW lasers, adequate stability can often be achieved even without bias (and hence feedback control). In this case, both the transmitter circuit and the package are simplified; for example, no photodetector is required in the package. NTT has just recently reported an all-plastic package for loop lasers, which is another major step toward an ultimate low-cost laser for FTTH (Fukuda et al. 1995).
Fujitsu modifies the laser chip to achieve relaxed tolerances on laser-fiber coupling by integrating a beam-expander region. This increases the mode-field diameter to 3.5 microns, reducing the sensitivity (dB/micron) to transverse laser/fiber displacements by a factor of three. Finally, packaged loop lasers must show high reliability at elevated temperatures, the usual goal being 105 hr at 85 degrees C.
Although Toshiba does not currently have a strong vertical-cavity surface-emitting laser (VCSEL) program, its researchers believe VCSELs are a candidate for FTTH lasers. They can be wafer probed, die-bonded more readily than edge emitters, and they can be coupled easily because of their symmetric far-field pattern. However, VCSELs are currently short-wavelength devices. When 1310 nm VCSEL technology eventually appears, reliability will need to be established and wide-temperature operation demonstrated or achieved through device design, etc.
In the area of loop lasers, the Japanese are currently ahead of efforts in the United States, although wide-temperature MQW designs also are in production in the United States (at AT&T and Lasertron, for example). Also, several companies have the capabilities in chip design, production, and packaging to move into this area. The issue becomes one of market size and profit: with FTTC, two lasers are needed for every 4 to 16 homes, on the average. For the entire United States, this is on the order of 30 million lasers, total, including replacements, growth, etc. This is roughly one year's production of packaged CD lasers for a company like Sony. (The packaged devices are of similar complexity; only the optical coupling to single-mode fiber is unique to the loop laser.) If the price reaches $30 as expected (Endnote 1) , this is less than a billion dollars in sales spread out over 20 years. Although this is a significant business, it will be divided among many suppliers, and the Japanese are positioned to capture most of it. Key factors that could expand the number of required loop lasers are (a) the potential for worldwide sales, (b) other applications such as high-speed data links, and (c) a shift back to FTTH, which multiplies the market for loop lasers and detectors by roughly an order of magnitude.