By Stephen Hardy
The wait for economical 100-Gbps systems should give 40-Gbps technology plenty of time to take root in carrier networks, analysts agree. The metro should see significant 40-Gbps deployments begin this year.
Ciena Corp.’s announcement in May that it will supply 100-Gbps-enabled versions of its CN 4200 RS FlexSelect Advanced Services platform to NYSE Euronext would seem to herald the start of the 100-Gbps networking era. However, analysts consulted for a recent Lightwave webcast on the 40- and 100-Gbps market say there is actually no time like the present for 40-Gbps technology. What may sometimes appear to be the unloved stepchild of high-speed networking should transition from a means of router interconnect to a viable option for metro networks this year, they say.
It’s common to focus on the leading edge of technology when looking at industry trends, says Vladimir Kozlov, founder and chief executive officer of the transceiver market research and consulting firm LightCounting (www.lightcounting.com). However, there’s a long tail for previous technologies, and the bulk of the market is often much slower to adopt newer technologies than well-known early adopters. This axiom certainly applies to comparisons of the prospects for 40 and 100 Gbps. “The bulk of the market is much more inert than the leading edge,” he says. “And if 40G gets into the bulk of the market—and we believe it will—it will stay there for years.”
According to LightCounting analyst Roy Rubenstein, 100-Gigabit Ethernet module sales should begin ramping next year, but line-side 100-Gbps applications will not follow suit until 2012. Kozlov agrees, saying that his firm projects perhaps 3,000 ports of 100-Gbps shipping that year, versus about 50,000 for 40 Gbps.
Ron Kline, who follows optical systems sales in his role as research director, optical networks at Ovum (www.ovum.com), and Michael Howard, principal analyst and cofounder of Infonetics Research (www.infonetics.com), say that a growing number of these ports will be found on metro equipment, particularly 4×10-Gbps muxponders (see figure). For this reason, the prospects for 40 Gbps look good even in the face of upcoming 100-Gbps systems, they assert.
“I believe it has a significant role to play,” says Kline. “While it looks like it’s niche now, the larger carriers I don’t believe look at it that way.”
One reason for the use of 40 Gbps in the metro is that rising bandwidth demands have finally met declining technology costs. Such offerings still haven’t met the magic “2.5× the cost for 4× the capacity” versus 10 Gbps. However, that has more to do with the continuing price declines for the lower-speed systems rather than any failings of 40-Gbps development, Kline and Howard explain.
Kline says the average cost for a 40-Gbps line card currently hovers between $50,000 and $55,000—4× to 6× current 10-Gbps prices. (Howard puts 40-Gbps prices at 5× to 7× the current cost of 10-Gbps technology.) However, that price tag is still low enough to make 40-Gbps upgrades to existing networks economically advantageous when compared with laying new fiber or installing entirely new systems on routes with capacity exhaust.
Festoon and unrepeatered submarine networks also will find 40 Gbps an attractive option, he adds.
What form will 40G take?
A transition from the first generation of 40-Gbps systems—marked primarily by line cards and systems using either phase-shift binary transport (PSBT)/optical duobinary or carrier-suppressed RZ—to the current second generation has helped the reduction in price. The current generation is module-based and uses some form (likely proprietary) of differential phase-shift keying (DPSK).
Kozlov and Rubenstein believe DPSK technology will dominate sales over the next two or three years. However, the evolution of 40-Gbps modulation formats hasn’t ended. Modules based on differential quadrature PSK (DQPSK) have begun to appear, promising better performance if somewhat higher cost. Kozlov says this format should prove particularly popular in Asia. (Vendor sources say the format also should find favor in North America if it can be made cost-competitive.)
The analysts also predict that once the supplier community perfects the dual-polarization QPSK the Optical Internetworking Forum has chosen as the focus of its work at 100 Gbps, modules of this type also should enter the 40-Gbps market. That means that while carriers should find 40 Gbps an increasingly economical option for upgrading 10-Gbps networks, the technology underpinning this alternative will continue to change.
Stepehn Hardy is editorial director and associate publisher of Lightwave.
Lightwave Webcast: 40G/100G Market Trends
Lightwave Webcast: 40G/100G Network Technology and Implementation
Lightwave Online: CIR Predicts First Revenues for 40/100GbE by 2010