OptiMight Details Long-Haul Box

OptiMight Communications Inc., a startup founded by Wu-Fu Chen and physicist Ilya Fishman, is set to unveil the first offering in its three-product roster next week at OFC. Its product, the OMC 1600, an optical transport box for long-haul networks, purports to be able to send up to 1.6 terabits of data per second -- 160 channels at 10 gigabits per second -- over any type of fiber without using Raman amplification.

Raman amplifiers are used to strengthen optical signals so that service providers don't have to use as much expensive electrical regeneration equipment along network routes. OptiMight says its OMC 1600 can transmit signals up to 4,000 kilometers without needing signal regeneration or Raman amplification.

Networks using OptiMight's boxes will, however, need to use erbium doped fiber amps (EDFAs) every 80 to 100 kilometers and signal conditioners every 500 kilometers in between the transmission terminals at either end of the long-haul connection (see Optical Taxonomy and The Ultimate Backbone ).

The technology behind OptiMight, dubbed full-spectrum WDM, is a trick that alters optical signals in both the time and frequency domains. In the time domain, it uses the conventional WDM (wavelength-division multiplexing) technique of quickly encoding each OC192 (10 Gbit/s) data stream together onto the optical signal.

But OptiMight also applies additional frequency modulation to the signal so that each channel's spectral width is broader and can carry data farther. The result is that OptiMight is able to fill 30 GHz within each channel on the International Telecommunication Union (ITU) WDM grid. Other WDM systems usually use 10 GHz of spectrum per channel.

Where this matters most is in network economics, says Clarel Thevenot, OptiMight’s director of marketing. "Our solution will save carriers the cost and maintenance associated with all the Raman amplifiers and extra regeneration stations that would otherwise be needed." (See OptiMight Reaches for Distance.)

"The goal is to allow carriers to use the same product platform when connecting Washington D.C. to Philadelphia as they'd need to connect Washington D.C. to Chicago," he says.

But skeptics point out that although OptiMight’s box may be used for both long- and ultra long-haul connections, it's walking into a market where Williams Communications Group (NYSE: WCG) and Broadwing Communications (NYSE: BRW) are using Raman-based networking technology from Corvis Corp. (Nasdaq: CORV); and Genuity Inc. (Nasdaq: GENU) is using Nortel's Raman-based long-haul gear. The implication: If Raman costs and maintenance were a disadvantage, carriers wouldn't be going ahead with such field trials and purchases (see Corvis Goes Coast to Coast).

More to the point: The market is getting more crowded by the day. Besides Corvis and Nortel, you'll find Alcatel SA (NYSE: ALA; Paris: CGEP:PA), Ciena Corp. (Nasdaq: CIEN), PhotonEx Corp., Solinet Systems Inc., Solstis, and Sycamore Networks Inc. (Nasdaq: SCMR) all angling for part of the long-haul market. Then there's the argument that, by not using Raman amplification, OptiMight is giving up bandwidth in order to compete on distance. "They seem to be painting themselves into a corner of being able to handle long, skinny traffic routes," says one equipment provider familiar with OptiMight's product.

OptiMight’s Thevenot counters that the overall throughput of OptiMight’s system is the same as anyone else's. "They'll say we can't pack as many channels because we're using a wider spectrum per band, but, from a total capacity standpoint, we're exactly the same," he says.

OptiMight’s technology also allows for bi-directional transmission over fiber, a characteristic that has arguable advantages. On the one hand, bi-directional transmission makes more efficient use of fiber that's already in the ground. On the other hand, early WDM efforts proved that when you have go- and return-traffic on the same fiber, the operational issues can be sticky because it becomes harder to pinpoint network problems.

Anyway, OptiMight can make an interesting case for cutting carriers' costs in a variety of ways. Its marketing materials say it can save carriers more than 60 percent over the equipment and operating costs of existing solutions. While that's probably an elastic fact, it is interesting that OptiMight’s box can be used with any kind of fiber.

In the end, the OMC 1600's success will depend on how it plays in the field. OptiMight says it's completed lab trials and is moving on to field trials with "two large IXCs." The company should be realizing revenues on those boxes by early next year.

OptiMight plans to have the OMC 3200, its OC768 (40 Gbit/s) transport box, ready for general availability by the third quarter of of 2002. That one will be followed by the 6.4-Tbit/s OMC 6400 box, expected in late 2003.

-- Phil Harvey, senior editor, Light Reading http://www.lightreading.com