Impressive growth surge predicted for optoelectronic subsystems

Impressive growth surge predicted for optoelectronic subsystems

Market sales for fiber-optic transmitters, receivers and optical amplifiers are forecast to escalate from $1.01 billion in 1994 to $5.84 billion in 2004

SABA HAILU

ELECTRONICAST CORP.

North American production of fiber-optic optoelectronic subsystems--transmitters, receivers and optical amplifiers--is expected to rise sharply over the next decade, from $1.01 billion in 1994 to $2.8 billion in 1999 and to $5.84 billion in 2004. The average annual growth rate of the subsystem market is calculated at 23% during the first five years, and at 16% the second five years.

In fiber-optic subsystem applications, telecommunications outdistanced premises data networks, cable-TV and broadcast, military and aerospace, and specialty applications in 1994 with 77% of the market share, at $776 million. However, premises data communications applications are estimated to grow impressively from a 9% market share in 1994 to 24%, or $627 million through 1999. All forecasts are based on the complete, final value of functional transmitters, receivers and optical amplifiers.

Historically, the fiber-optic optoelectronic subsystem market has been dominated by laser-diode-based transmitter and receiver sets or data links. This domination emerged because the early deployment of fiber-optic applications consisted mainly of long-distance, high-data-rate installations whose performance requirements could not be met by light-emitting diodes, or LEDs. In addition, optoelectronic subsystem costs were small compared to the total system cost, and little incentive existed for cost reduction.

Presently, network installations are moving toward the use of shorter optical cable lengths and moderate data rates, especially in premises applications. Consequently, the interconnecting cable represents an influential contributor to the total installed system cost. In these installations, the cost of fiber-optic cable competes directly with coaxial cable and twisted-pair copper wire costs. Furthermore, the high cost of fiber-optic data links has restricted market penetration into premises networks.

However, technology improvements have enabled LEDs to achieve high production volumes and correspondingly low chip costs. Over the next few years, LEDs are predicted to gain a substantial share of the low-to-moderate date-rate range: up to OC-12, or 622 megabits per second. This data-rate segment is figured to become a high-volume market. The operational performance of LEDs is therefore being aggressively improved. For example, LEDs that accommodate longer wavelengths to match the characteristics of low-loss fiber are now available.

Laser-diode costs are also decreasing because of increased production and increasingly relaxed performance requirements for short-distance applications. Another possible cost savings for transmitter and receiver data links is anticipated with the implementation of massively parallel optical interconnects for inter- and intraenclosures.

In 1994, laser-diode-based transmitters and receivers grabbed a 72% share, or $723 million, of the North American market. With a 70% market share anticipated during the next decade, these subsystems are expected to dominate this market sector and become a $4.1-billion market in 2004.

At a predicted average market share of 19% over the next decade, LED-based transmitters and receivers are also expected to surge in sales from $168 million in 1994 to $642 million in 1999 and to $1.14 billion in 2004.

Telecom applications

The sales growth of fiber-optic optoelectronic subsystems for telecommunications applications over the next decade is projected to soar, based on the expected burgeoning use of synchronous optical network-rated data links in interexchange and interoffice trunk lines and in feeder lines. This rapidly exploding market is predicted to surge at a 20% average annual growth rate from 1994 to 1999, and to climb at a 25% growth rate from 1999 to 2004. During that time, market sales are estimated to rise sharply from $320 million in 1994 to $794 million in 1999 and to $2.47 billion in 2004.

Because of the projected domination by Sonet-rated transmitters and receivers, market sales of conventional transmitter and receiver data links are expected to drop dramatically from $348 million and a market share of 45% in 1994, to $84 million and an 8% share in 1999 and to $24 million and a 1% share in 2004.

Another market dynamic for evaluating telecommunications optoelectronic subsystems considers the amount of fiber-optic cable deployed (in kilometers) by interexchange carriers versus the number of transmitters installed. Analyzing the available data results in a numerical ratio value of 7.6 for the year 1994 and 6.9 in 1999. The slight decrease in ratio value over time is attributed to the eventual completion of several large optical networks in North America.

The number of installed transmitters should decrease more slowly than the amount of deployed fiber-optic cable because faster devices are steadily being incorporated to add network capacity without the need to install more cable. Other market factors, such as the implementation of wavelength-division multiplexing and the decreasing number of dark fibers, also contribute in decreasing the ratio of fiber cable deployed to number of transmitters installed.

From 1999 to 2004, however, the ratio is expected to zoom to 17. This rapid climb is expected because network link lengths--the distance between repeaters-- is expected to increase markedly in interexchange networks.

Another market dynamic deals with the trend in the average link length installed by local exchange carriers. These carriers are projected to experience a decrease in average fiber-link length from 8.4 km in 1994 to 6.5 km in 1999. This decline is expected to continue through 2004, when an average link length of 2.8 km is estimated.

This steady decline is predicted because of the increasing percentage of fiber-optic cables being deployed in the distribution segment of optical networks. The shift in cable deployment from long interoffice lines to short distribution lines is expected to shorten the average link length for local exchange carrier applications.

Premises data networks

Data communications networks are rapidly moving toward faster data rates per terminal; a higher number of terminals served per node, hub or concentrator; a higher number of terminals per local area network, and a higher number of interconnected local area networks. Another market-driving force deals with user needs to interconnect corporate enterprise networks to global broadband networks. For these reasons, the deployment of fiber distributed data interface networks, 100-Mbit/sec Ethernet networks, Fibre Channel networks and asynchronous transfer mode/Sonet switching/transport networks is figured to multiply dramatically during the next decade.

Cable-TV networks

Fiber-optic network deployments by cable-TV operators expanded by more than 50% during 1994. Sales of transmitter/receiver sets and optical amplifiers for cable-TV network applications are predicted to grow from $51 million in 1994 to $439 million in 1999, at an average annual growth rate of 54%. A drastic slowdown--4% growth from 1999 to 2004--resulting from network saturation should confine this market to $539 million.

Regional Bell operating companies and independent telephone operators are planning the delivery of broadband services to residential customers, although the cost issues of providing multiple interactive television services have not yet been resolved.

Notwithstanding, major cable operators--the multiple system operators--foresee financial advantages in moving aggressively in increasing bandwidth, reliability and bidirectional capabilities of their fiber-optic networks. Operating more than 90% of the available optical cable-TV networks, the operators have become attractive acquisitions or partners for the Bells. On the other hand, the multiple system operators may choose to remain independent and compete head-on with the telephone companies.

Military and aerospace

Sales of fiber-optic optoelectronic subsystems for military and aerospace networks and equipment are expected to nearly triple from $53 million in 1994 to $143 million in 1999 at an average growth rate of 22%. This growth is expected to be maintained at a 19% rate to 2004 and produce a $342 million market. Tightening of military and aerospace spending is expected to continue unabated during the next decade. However, the major part of this spending is gauged to focus on upgrading existing fiber-optic networks, systems and equipment.

Specialty subsystems

The specialty applications market is presently dominated by intramachine (inside-the-box) fiber-optic links. Specialty optoelectronic subsystem sales accrued $35 million in 1994, but is calculated to rocket to $471 million in 1994 at an average annual growth rate of 68%. From 1999 to 2004, sales are expected to expand 13% and reach $869 million.

Nearly all of the optoelectronic subsystems used in specialty network applications are bought by original equipment manufacturers. A minor specialty market does exist, though, for upgrading subsystems. The applications covered by specialty networks include:

Inter- and intramachine connections

Automotive and vehicular signal interconnect harnesses

Fiber-optic and other optical and optoelectronic instrumentation

Medical diagnostic equipment and fiber-optic sensors

Other miscellaneous uses.

Technology trends

Historically, lightwave power for optical-fiber transmission has been supplied by either LEDs or semiconductor laser diodes. However, to meet the current increased demands for longer span lengths between repeaters and to achieve high split ratios to subscribers, network planners have been pushing for higher lightwave power levels coupled into the fiber. In the past, lightwave powers of 10 to 13 decibels relative to milliwatt were considered adequate. Presently, however, network planners are specifying 18 to 21 dBm. The upper level of 21 dBm has been laboratory demonstrated, and users are clamoring for even higher power levels.

Alternate optical-power solutions being investigated include: high-power laser diodes, both single diode and diode arrays; optical fiber amplifiers, from single pump to quadruple pump; and solid-state (crystal) lasers.

A technology that is similar to that of optical fiber amplifiers is being employed by pumped, doped-fiber light sources. Substantial research and development have been targeted at these light sources, and some pumped doped-fiber light sources have been demonstrated in certain applications.

Optoelectronic integration

The major cost pressure related to lowering costs of broadband network connections to homes is driving system designers toward transmitter/receiver/splitter integration. Companies such as AT&T, AMP Inc. and Hewlett-Packard Co. are aggressively pushing fiber-in-the-loop subsystem integration.

For example, parallel optical interconnect transmitter and receiver units represent a major area of next-generation integration. These units integrate more than 30 laser diodes, photodiodes and electronic integrated circuits into a 1-cubic-inch package. Most developments have concentrated on designing vertical cavity emitting laser diodes. For telecommunications applications, which mandate 5- to 10-gigabit-per-second transmission and high spectral stability requirements, external modulators are expected to be integrated with laser diodes.

Other investigations are exploring the integration of optical amplifiers with transmitters (booster amplifiers) and receivers (preamplifiers). As deployment volumes increase, fiber-in-the-loop networks expand and optoelectronic device reliability continues to improve, subsystem prices are expected to drop and optoelectronic integration is forecast to climb steadily with widespread fiber-optic network deployments during the next decade. u

Saba Hailu is a senior research analyst in the fiber optics group at Electronicast Corp. in San Mateo, CA.