When the market's a little crazy, it helps to look at statistics. From RHK's 2/00 report:
10Gbps termination components grew 100 percent to $180 million in 1999 and will reach $2.0 billion by 2003.
Revenue for 10 Gbps termination components rose at a 100 percent rate in 1999, from $90 million to over $180 million. From 1999 to 2000, the market for these components will grow 130 percent to $430 million. By 2003, revenue will reach $2.0 billion, which corresponds with a CAGR of 80 percent between 1999 and 2003. The most significant growth for 10 Gbps components will come in 2000, when units will quadruple and revenue will triple. Contributing to this growth is the continuing increase in network traffic. Increasingly, service providers are planning to use 10 Gbps, DWDM systems and are evaluating network edge routers with 10 Gbps interfaces. A greater number of DWDM system manufacturers have 10 Gbps systems ready. The number of component suppliers with 10 Gbps capability is steadily growing. This coordination, from service providers through their supply chains to component manufacturers, is a strong sign of a market prepared for high growth. The increasing preference of service providers for 10 Gbps DWDM wavelengths over 2.5 Gbps channels is depicted in Figure 1-8, which shows that 10 Gbps components will garner more than 50 percent of termination component revenue by 2003. [emphasis mine]
. . .
Over the next few years, InP modulated sources will grow in market share relative to externally modulated sources, as has occurred with 2.5 Gbps modulated source components.
Pumps grew 65 percent to $245 million in 1999 and will reach $1.2 billion by 2003
Revenue for pump laser modules used in terrestrial DWDM EDFA gain blocks is increasing 65 percent in 1999, from $145 million to $240 million. By 2003, revenue will increase to $1.2 billion, which corresponds with a 50 percent CAGR from 1999.
Figure 1-12 shows that 980 nm pumps make up 75 percent of the $240 million 1999 pump market. Since the mid-1990s, when manufacturers demonstrated sufficient reliability for 980 nm pump modules, gain block designers have opted increasingly for 980 nm over 1480 nm pumps. The preference for 980 nm pumps stems from their superior noise and thermal characteristics.
Further examination of Figure 1-12 shows the greatest pump growth in 1480 nm pumps and 980 nm pumps with kink-free powers of at least 150 mW. High-power 980 nm and 1480 nm pumps are growing the most rapidly, as gain block designers are adopting new architectures for amplifying increasingly greater numbers of photons. Revenue for 1480 nm pumps is growing 75 percent to over $50 million; high-power 980 nm pump revenue is growing 150 percent to $90 million. The change in architecture will be discussed below.
The high-power 980 nm pump growth is even more impressive than the 150 percent increase, as the growth also incorporates a major change in product specification and construction. In 1999, gain block designers placed a stronger emphasis on grating stabilization of 980 nm pumps. The stabilization requirement arose because a new generation of high-channel-count EDFAs needed to operate under conditions that increased the difficulty of satisfying gain flatness requirements. These conditions include operation of the EDFA to the lower end of the 1530-1565 nm C-band and changing the channel loading of EDFAs during their service life. When the number of DWDM channels being amplified changes, the power required from the pump correspondingly changes. When the pump does not have grating stabilization, this change in pump power leads the pump to operate at a slightly different wavelength. The change in wavelength causes the gain flatness of the gain block to vary. To maintain gain flatness, gain block manufacturers have required 980 nm pump modules to include grating stabilization. The incorporation of gain stabilization, along with 150 percent revenue and 200 percent unit growth, makes the growth in 980 nm pumps of at least 150 mW more noteworthy.
Beginning in 2000, shipments for 1480 nm pumps will increase faster in units and revenue than for 980 nm pumps. Shipments of 1480 nm pumps will rise from 1999 to 2003 at a 90 percent CAGR for units and a 70 percent CAGR for revenue, to $500 million. Shipments of 980 nm pumps will grow at a 50 percent CAGR for units and a 40 percent CAGR for revenue, to $700 million, over the same period.
The number of units shipped will grow faster for 1480 nm pumps than for 980 nm pumps because designs for newer, higher-channel-count and higher-bit-rate gain block increasingly emphasize cost per DWDM photon. As shown in Figure 1-13, gain blocks currently rely much more on 980 nm pumps than on 1480nm pumps. In gain blocvks that typically contain two pumps, the designer uses 980 nm for its superior noise performance. Noise performance is especially important in the first erbium doped fiber of a gain block, where amplification of noise is strongest.
As the number of DWDM channels grows and as DWDM bit rates rise, need for gain blocks with greater pump power will intensify. Gain block designers will obtain this additional pump power in two ways. They will purchase pumps with output powers, as indicated in Table 1-8. They will also include more pump units per gain block. As show in Figure 1-13, the number of pumps per gain block will more than double from 2.0 in 1998 to 4.2 by 2002.
Most of this increase in the number of pumps per gain block will be attributable to 1480 nm pumps. The number of 1480 nm pumps per gain block will rise from approximately 0.5 in 1998 to 2.2 in 2003, whereas the number of 980 nm pumps will rise from approximately 1.5 to just 2.0 over the same period. The faster increase in the number of 1480 nm pumps per gain block will occur because noise performance is no longer as critical after the initial pump. Instead, in these additional pumps, cost per DWDM photon is a greater concern. Since the cost of generating DWDM photons is less expensive with a 1480 nm pump, the demand for more pump power in gain blocks contributes to a higher 1480 nm pump growth rate.
The slower growth rate for all 980 nm pumps masks the high growth of the 980 nm pumps of at least 150 mW. Figure 1-14 shows that the low-power 980 nm pumps peak in 2001 and then decline. The figure also makes clear that the high-power 980 pump segment exceeds the entire 1480 nm pump market. The high-power 980 pump segment will rise to $670 million by 2003, a CAGR of 65 percent from 1999. In addition, the average selling prices of high-power 980 nm pumps will stay higher than the average selling prices of 1480 nm pumps, as can be seen in Table 1-8. For example, in 2003, the average price of a high-power 980 nm pump will be just under $1,400, whereas the average price of all the 1480 nm pumps will be in the mid-$800 range. RHK expects 980 nm pump prices to remain higher because making high-power, high-reliability pump modules is more difficult at 980 nm than at 1480 nm.
>>>>>>
According to friends in the industry, since this report was published, the numbers have been increased. However, I don't have those figures so can't verify.
Pat |
