DT deal doesn’t mean 40G’s case is closed

As described in last month’s Analyst Corner (see “40G optical networks: why, how, and when?” by Tom Hausken, February 2005, p. 39), several demonstrations of 40-Gbit/sec networking have taken place over the past two years, so it seemed inevitable that the implementation that had been “just around the corner” for the past three years was at hand. David Dunphy, principal analyst, optical infrastructure, at Current Analysis (Sterling, VA) says that DT should be considered something of an early adopter. Nevertheless, “the impact that Deutsche Telekom has is essentially you’ve got a fairly solid well-known PTT that’s relatively conservative going out and adopting the technology. That certainly makes it easier for me to stand up in another service provider and say I think it’s time for us to do it too,” he adds.

However, few such people are likely to stand up this year, it seems. If one believes that the deployment of a new technology requires a confluence of customer requirements, technological maturity, and an economical price point, 40-Gbit/sec systems have work to do in all three areas.

“From a capacity requirement standpoint, we’re not at a point right now where most carriers are going to feel compelled to run out and install 40G based on the current traffic volume,” Dunphy says on the requirement end. Many network upgrades in the near future are likely to involve moving from 2.5 to 10 Gbits/sec, rather than 10 to 40 Gbits/sec, due to the large number of 2.5-Gbit/sec links now in deployment. Many 10-Gbit/sec systems in the field still have a lot of available wavelengths, Dunphy believes. “Even if you might have started to build 40G now, considering what you already had in the ground in available capacity at 10G, why would you?” he asks.

One reason might be the deployment of the latest generation of core packet routers that carry 40-Gbit/sec interfaces, such as the CRS-1 from Cisco Systems. Dunphy agrees such systems will drive 40G deployments, and Stefan Kindt, managing director of Marconi Communications GmbH (Backnang, Germany), asserts the economics of 40G router interfaces provided DT with a major incentive to insist on 40G DWDM equipment. The carrier studied the 10G versus 40G question extensively, Kindt says. “I think that the results of that calculation must have been so convincing that they decided that whatever extension to their existing core network they would be building would have to be with a platform capable of supporting and evolving to multiple 40G transmission,” he reasons.

The advent of 40G transponders that fit within standard rack slots also influenced DT, Kindt says. Rather than the “pizza box” approach of the previous generation of high-speed subsystems, Marconi co-developed with CoreOptics GmbH (Nuremburg, Germany) a 40G transponder that fit directly into the Multihaul 3000. The transponder form factor represents the third generation of 40G subsystems, according to Saeid Aramideh, vice president of marketing, Americas, at CoreOptics. This generation can be integrated from an OAM&P perspective into the existing system, Aramideh asserts. “It’s a pluggable system. It obviously meets the performance specification of DT and Marconi and such. And from a cost perspective, we think it has a lot of promise. It’s still not at 2.5 times the 10-gig cost that obviously you’re shooting for. But it is within striking distance,” he says.

While such transponders represent an important advance for the technology leg of the triad mentioned above (other subsystems vendors such as Mintera and Stratalight Communications have pushed the same approach), they don’t remove all of the physical challenges to 40-Gbit/sec deployment. “I feel that we will see an evolution of 40G interfaces becoming more and more tolerant to fiber impairment, particularly PMD [polarization-mode dispersion],” Kindt predicts, highlighting a major focus of current research. “I think that once you reach a certain threshold of PMD tolerance in your transponders and you have 40G transmission over decent distances-not just on selected fibers but on the general fiber population out there-that will really be the point where 40G will become the norm out there in the transmission networks.”

Such an achievement will enable 40G to reach the cost target to which Aramideh alluded. “If you look at why 40-gig was really expensive, there were two elements,” he explains. “One element was the cost of the node, especially because everything was tailor-made. And the second thing was the line, and the cost of the line was extremely expensive because of all of the [dispersion compensation modules] and tunable dispersion compensators that you had to put on the line.”

Aramideh believes that more cooperation among subsystems and components vendors, particularly in the form of additional multisource agreements, would go a long way toward making 40G more economical. “Obviously, not everything can be based on a standard, such as clock and data recovery schemes or ultra-FEC [forward error correction] at 40-gig, which is on the leading edge of what’s available out there. But we’d welcome standardization activities or efforts,” he says.

Advancements in requirements, technological maturity, and costs have made it more likely that 40G technology will reach the field, as the DT contract proves. “Our wish is that more announcements will follow,” Aramideh concedes. “We definitely have been seeing a lot of excitement about 40-gig in China for shorter-haul applications. We have seen some progress in North America as well.”

Stephen Hardy is the editorial director and associate publisher of Lightwave.