Earlier this week, Tropic Networks Inc. announced details of its reconfigurable optical add/drop multiplexer (ROADM), a key part of its plans to help carriers drive down the capital and operating costs of metro DWDM networks (see Tropic Debuts Managed ROADM).

The actual announcement isn't exactly earth shattering, because the ROADM won't be available until the third quarter of this year, and a non-reconfigurable, static version of it has been in trials with various carriers for a while.

Still, much of Tropic's business case is bound up in enabling carriers to reconfigure their metro networks from a remote management console, and it’s worth taking a closer at this. It’s also worth comparing Tropic with two other vendors promoting the same idea:

Marconi Corp. plc (Nasdaq/London: MONI), which has been shipping a ROADM for more than a couple of years. Like Tropic’s product, Marconi's PMA32 is a pure optical device to which any sort of equipment -- such as a router, Ethernet switch, or Sonet/SDH add/drop mux -- could be connected. Marconi also bolts on SDH sub-wavelength multiplexing in some instances, making the whole thing comparable to developments at...

Photuris Inc.: This startup has combined a ROADM with a Sonet ADM (see Photuris Metro Box May Draw $40M and Photuris Gets New CEO and $40M).

Tropic’s big selling proposition is that it can solve the main problem that’s prevented metro DWDM from being widely deployed -- i.e., more often than not, the return on investment for the technology doesn't stack up for carriers.

There are two reasons for this, according to Rob Lane, Tropic's VP of marketing and business development. First, the gear costs too much, and second, operating costs are way too high. That's because customer requirements change frequently, and every time the network has to be reconfigured, highly paid technical staff have to effectively redesign portions of it and visit sites to rebalance the optics by making manual adjustments to equipment.

This second issue is the real killer, according to Tropic. Lane maintains that there's no shortage of funds to build metro networks. "Some of the RBOCs are spending upwards of $300 million a year," he says. But the business case for using this cash to build metro DWDM networks often doesn't stand up. Tropic cites a study published by McKinsey & Co. in December 2002 that says carriers need to cut operating expenses by 34 percent in order to achieve a 12 percent return on investment.

Capex

Tropic tackles the first issue, reducing capital expenditure, by eliminating a lot of transponders in its equipment, transponders representing a big slice of the overall cost of an average DWDM network.

It does this by getting the carriers, or their customers, to fit International Telecommunication Union (ITU) Grid optics in attached equipment. This means that the wavelength of signals doesn't need to be changed as they enter or exit the DWDM system, and that eliminates a transponder in each case. In a complete network, this can result in a 65 percent reduction in transponders, according to Tropic.

It’s worth pointing out that ITU Grid optics, based on singlemode fiber, can cost up to four times as much as the usual types of optics used in customer premises equipment, which use multimode fiber. Also, carriers would have to fit a different line card to Tropic’s ROADM depending on what type of equipment is being attached, so a site visit by an engineer still seems inevitable, even if the engineer has a fairly mundane task to accomplish.

Ashish Vengsarkar, VP of product marketing and a founder of Photuris, says carriers also have reservations about getting customers to fit special ITU Grid optics to their gear. First, picking the right module for the attached equipment isn't easy: "There's a lot of optical issues to consider," he says. Second, messing about like this muddies the waters when something goes wrong and the finger-pointing starts.

Photuris sidesteps this whole issue by integrating its ROADM with a Sonet ADM, eliminating transponders. As noted, Marconi has added sub-wavelength multiplexing to the PMA32 so that it achieves something similar. “This isn’t anything new” says Robin Abel, director of product marketing and development for Marconi’s metro products. Offering sub-wavelength multiplexing is a growing requirement of carriers, he notes.

There’s always a trade-off between capex and opex with network equipment, according to Abel. The PMA32 costs “slightly more” than Marconi’s equivalent static product but offers higher opex savings. One likely reason the PMA32 costs more is that it needs a fancy liquid crystal optical backplane, supplied by Corning Inc. (NYSE: GLW).

Opex

As noted, Tropic maintains that the real key to improving return on investment is reducing carriers' operating expenditures.

In this respect, Tropic and Photuris sing the same tune. Both companies promote reconfigurable networks as the solution, and both make the point that in order to reconfigure networks remotely, the optical layer needs to be managed. Specifically, the power of each wavelength needs to be monitored and adjusted each time a new service is set up, to avoid knocking everything out of balance. Power also needs to be monitored to spot incipient problems.

Tropic has a special way of monitoring power which involves adding a sort of bar code to each wavelength (see Tropic Touts Tracking Approach). Engineers back at base can adjust the power at any point in the network using variable optical attenuators.

Vengsarkar says Photuris has an "inhouse innovation" which achieves similar results. This involves automatic dynamic power equalization of wavelengths, and optical power monitoring.

Marconi’s Abel says reducing opex is just one of the reasons that carriers buy the PMA32. Another is being to able to provision services quickly. “The whole principle is that you plug in the tributary transponder, you fire it up, and you walk away. You don’t have to spend an age balancing the ring.”

So what’s there to learn from Marconi's experience with the PMA32?

First, reconfigurable metro DWDM networks work. “We’ve had the product in PTT networks carrying live traffic for more than two years,” says Abel. However, he concedes that Marconi had some early production problems with the PMA32, which delayed deliveries for a while. This raises the question of whether Tropic and Photuris will encounter similar problems when and if they go into commercial production.

Second, Marconi has had mixed success at getting carriers interested in reconfigurable networks, although perhaps this will change, now that reducing opex has become a priority for many of them.

Marconi’s biggest success by far was in selling the PMA32 to British Telecommunications plc (BT) (NYSE: BTY; London: BTA). It’s a key part of its famed multibillion-dollar frame contract with BT (see Marconi wins $3+ Billion Contract). It’s also sold the PMA32 to other carriers in Europe and Australia, in some cases packaged as part of its SDH offering.

Still, Marconi isn’t expecting its ROADM to trigger strong growth of metro DWDM sales, judging by a recent Light Reading interview with CEO Mike Parton, who said: “My view is that the business case for metro DWDM is still not proven. People are putting in higher-bandwidth SDH, or Sonet in the U.S. The metro DWDM market is certainly not a very big market today, is it?” (See Mike Parton, Marconi.)

Tropic and Photuris are clearly hoping this isn't the case, and are racing each other to clinch their first customer contracts.

Photuris claims to be further ahead than Tropic, in that it's had its ROADM in trials for some time. This includes trials with an RBOC, according to Vengsarkar. Tropic, however, says its non-reconfigurable equipment has been available since last September and has undergone lab and field trials with "several" Tier 1 carriers, cable operators, and CLECs.

Both companies acknowledge that they need to partner with an incumbent equipment vendor to clinch a deal with a big carrier, and both say they're close to doing that. "We're in detailed discussions with some very large equipment vendors," says Lane -- almost exactly what Vengsarkar says.

— Peter Heywood, Founding Editor, Light Reading