ROADMs Could Boost Components

ORLANDO, Fla. -- NFOEC 2003 -- Could it finally be time? For more than two years, vendors have been talking about reconfigurable optical add/drop multiplexers (ROADMs). Now, carriers might finally be ready to start deploying.

If that happens, it could trigger a long-awaited showdown among the component and subsystems companies all vying for attention in this space -- not to mention companies offering systems-level ROADMs.

RBOCs and cable operators are said to be preparing requests for proposal (RFPs), possibly for release this fall, that include ROADMs in the network.

"SBC [Communications Inc.] has been talking about it for a while, and it's supposed to come out in October," says Doug Green, principal analyst of Bradam Group LLC (no Website). "I think there will be some shorter-term action with the MSOs [multiservice operators]." More generally, Green sees hope for RBOCs to install more optical equipment, as equipment vendors have shifted their attention from creating new features to lowering operating expenses.

Of the systems vendors, Tropic Networks Inc. and Photuris Inc. have recently gained attention for their potential in upcoming carrier RFPs (see Metro DWDM Action Heating Up, Metro DWDM Renaissance? , and Redundancy, How Low Do You Go?)

But a more interesting story in the ROADM market may be the components players. There are about ten systems vendors and easily more than 20 components players, with many components players offering subsystem-level ROADMs.

The ROADM is a more flexible alternative to the OADM, one that allows on-the-fly adjustments to the number of added or dropped wavelengths. Ideally, a ROADM also does a reality check on the intensity of other wavelengths, as adds or drops can mess up the power levels of other channels.

The components/subsystems side is particularly muddled, with offerings based on tunable filters, liquid crystals, or micro electromechanical systems (MEMS). The mania here is spurred by a conviction that ROADMs are inevitable. "Every systems vendor has decided to use these things in OADMs," says Clarel Thevenot, vice president of marketing for Arroyo Optics.

Green agrees: "In the long term, if what's being talked about with the RBOCs happens, it could be a big market, not a toy market."

But getting there remains a hurdle. Cost is the biggest problem, as ROADMs are more expensive than static OADM counterparts.

In an NFOEC panel session on metro ROADMs, MCI (Nasdaq: WCOEQ, MCWEQ) fellow John Fee noted that first-day installation costs of a technology can outweigh any long-term savings. Given that carriers are still under financial strain and scrutiny, he said that it's too difficult to get financial higher-ups to bless a purchase that will save money in, say, three to five years.

That's important because it turns out passive DWDM is cheaper than ROADMs for the first few wavelengths that get lit, says Fee. As long as that holds, ROADMs will be in trouble. Automation of the manufacturing process would help lower the cost, as would the creation of optical ICs, something Fee said would be a particular boon to the ROADM cause.

In the meantime, the product requirements keep ratcheting up. Carriers ask OEMs for more features, and OEMs turn to the components players to make them happen. This trickle-down effect is happening "faster than ever before," says Bill Thompson, vice president of marketing for Clarendon Photonics Inc., a components firm specializing in silicon waveguides.

This has led to multiple spins of Clarendon's ROADM chip, the latest being the CP-3344D, prototype samples of which were completed earlier this week. The chip includes a crossbar to switch any wavelength to any port. OEMs had asked for that function, and Clarendon began tweaking its design accordingly three months ago, Thompson says. More changes are on the way: a second generation of the part will add a variable optical attenuator (VOA) to control the intensity of added wavelengths.

Another factor that leads to product changes is competition. "We've had competitors come out of stealth, and they've had capabilities we needed to match," Thompson says. "Every couple of quarters, you get a surprise from another competitor." It's always been that way, he says, but the downturn has meant no revenues between revisions.

Clarendon officials are proud of their ability to churn product revisions, but they don't want this going on forever. "We can't keep turning around new generations of products without the revenue," says Dr. Pierre Villeneuve, Clarendon's founder and CEO.

So if carriers are ready for ROADM deployment, what are vendors doing about cost? At the components level, many are hoping to adopt semiconductor processes to lower the cost of manufacturing, as ROADM building still relies on manual processes.

Companies such as Aegis Broadband, which makes thin-film filters, design their products to be built using semiconductor manufacturing equipment. Others such as Clarendon go all-out, designing components made of silicon so they can use the proven and relatively cheap manufacturing available to chipmakers. "Silicon economics and the use of a CMOS [complementary metal-oxide semiconductor] fab process are the way these kinds of devices are going to get cost-reduced," Thompson says.

Still, better manufacturing is only one step. "You can't automate your way out of a high-cost product," says Matthias Wagner, Aegis's CEO.

A few companies had NFOEC announcements tied to ROADMs:

  • Arroyo has begun sampling its wavelength blocker, marketing VP Thevenot says. The startup has shifted product plans a couple of times since 1994 and has yet to ship anything in production. Even an earlier version of the wavelength blocker was held back because Corning Inc. (NYSE: GLW) announced its product first.

  • Avanex Corp. (Nasdaq: AVNX) didn't announce new ROADMs but did display the products inherited from the Alcatel SA (NYSE: ALA; Paris: CGEP:PA) and Corning photonics groups (see Avanex Deal Reshapes Sector). Corning's ROADM is based on liquid crystal technology, while Alcatel's centers on a planar photonic circuit.

  • A startup called Alliance Fiber Optic Products Inc. (AFOP) claims to have completed a prototype of an 80-channel ROADM subsystem based on micro electro-mechanical systems (MEMS). AFOP didn't announce the part officially, but R&D engineer Chun He managed to plug it during the Q&A for the panel session that featured Fee and Green (interesting combination, that). AFOP doesn't have a customer for the design, He says.

— Craig Matsumoto, Senior Editor, Light Reading
boston beans 12/4/2012 | 11:26:54 PM
re: ROADMs Could Boost Components I believe that is www.aegis-semi.com not www.aegisbroadband.com

go red soxs
Sign In