Optical components

Corning Chops Wavelength Blocker

Corning Inc. (NYSE: GLW) announced today that it is discontinuing its wavelength switch and wavelength blocker products -- a move that could wreak further devastation on what's left of the optical switch market.

Corning also is closing down its Fountain Valley, Calif., facility, cutting roughly 190 jobs and resulting in a restructuring charge of $20 million to $30 million against first-quarter earnings.

From Corning's point of view, the news is not unexpected. The company is fighting to reach profitability and has indicated that it is reconsidering its future in the photonics industry (see Corning: Profits on the Horizon). Earlier this month, company executives told investors to expect further winnowing of Corning's photonics product line, and they admit Corning may have to exit photonics altogether.

What does it mean for Corning's customers? Companies such as JDS Uniphase Corp. (Nasdaq: JDSU; Toronto: JDU) also make wavelength blocker components (see JDSU Parades Picky Part). But adapting an existing switch design to fit the part from another vendor could take time and money, which most systems vendors can ill afford right now. With the market so small right now, systems vendors may not consider the effort worthwhile, so Corning's move could trigger some product cancellations.

By optical switch standards, Corning's wavelength blocker appeared to be a success, having been chosen by a number of key vendors making all-optical switches and reconfigurable add/drop multiplexers. Both Ciena Corp. (Nasdaq: CIEN) and Marconi plc (Nasdaq/London: MONI) were announced customers for the part, and it's likely that Corvis Corp. (Nasdaq: CORV) was also using Corning components as the guts of its all-optical switch (see Ciena Cleans Up OADM).

Ciena, for one, claims to be prepared. "Rest assured, we have alternative sources for any components for which we were relying on Corning," says spokesman Glenn Jasper.

Marconi could not be reached.

The wavelength blocker, which is based on liquid crystals, works as a high-performance version of a dynamic gain equalizer, which can attenuate individual channels down to virtually zero. This way, a carrier can drop channels from a DWDM stream without having to demultiplex the signals first (see Dynamic Gain Equalizers Diversify).

The key benefit, say Corning officials, is that the overall loss of the optical switches is less than for other types of switch fabric such as micro-electro-mechanical systems (MEMS). And this is what led to the wavelength blocker's apparent popularity.

Corning says it shipped more units than all competitors combined. But it is canceling the product because the market is still small in absolute terms, and volume sales will take longer than expected to emerge.

"It's an advanced product in an industry that doesn't have a need for advanced products right now," says Corning spokesman Daniel Collins.

That's not necessarily good news for competitors such as LightConnect Inc. While it's nice to lose a competitor that was "clearly the incumbent," it's also sobering to be reminded that the market remains sparse, says Yves LeMaitre, LightConnect's vice president of marketing.

"We need a big established player -- thank god JDS is still on that list," he says.

LeMaitre contends that the MEMS approach is preferable anyway -- an unsurprising attitude, considering LightConnect sells MEMS-based optical components. With MEMS, the option of a multiport configuration is possible, where a 1xN device could both block and switch wavelengths among multiple intersecting rings.

Customers are asking about such a device for use in metro networks, LeMaitre says. The obvious question will be cost, as the MEMS approach will have to prove itself cheaper than discrete components.

— Pauline Rigby, Senior Editor, and Craig Matsumoto, Senior Editor, Light Reading
deer_in_the_light 12/5/2012 | 12:38:41 AM
re: Corning Chops Wavelength Blocker JDS ?
alcabash 12/5/2012 | 12:38:41 AM
re: Corning Chops Wavelength Blocker I don't think so but can someone confirm ?
redface 12/5/2012 | 12:38:40 AM
re: Corning Chops Wavelength Blocker Dielectric filter based add/drop filters can be cascaded to drop multiple wavelength channels and allowing the remaining channels to go through. I wonder whether WB is accomplished by simply replacing the fixed add/drop filters with tunable add/drop filters which are based on very narrow band Fabry-Perot tunable cavity. Tunability can be achieved by filling the cavities with liquid crystal.

Can someone enlighten us?
whyiswhy 12/5/2012 | 12:38:39 AM
re: Corning Chops Wavelength Blocker Ha!

Wait till they get JDSU's revised price and delivery schedule...

jeb_knucklehead 12/5/2012 | 12:38:38 AM
re: Corning Chops Wavelength Blocker

I actually interviewed for a job with the group at Corning APT about 3 years ago. The engineering team was top notch. The design was based around a GRISM (a grating combined with a PRISM to increase the dispersion) and a nematic liquid crystal linear array . This device was pretty impressive for late 1999 early 2000 when I saw it.

The project was run by an old silver back gorilla. This guy had been 30-years at Corning, and knew nothing about optics. He was a total jackass. Corning actually relocated this guy from Corning, NY to "manage the project."

When I was there it was obvious as an interview candidate that there was significant tension between this guy and the lead engineer (who basically designed the whole thing by himself). I wouldn't be surprised the poor management killed this product and Corning APT in general.

What a shame and waste of a lot of really talented people.
lzrjck 12/5/2012 | 12:38:32 AM
re: Corning Chops Wavelength Blocker Amazing how badly LR screws this stuff up. Spectraswitch? Their device is NOT wavelength-selective! It wouldn't be so hard to find out who makes a similar INTEGRATED WAVELENGTH SWITCH. For those interested in accurate info, the short list is as follows:
Corning, JDS, Network Photonics, SLM, LightConnect, PolyChromix, Avanex, and some others who will probably die before they get product to market.

MEMS will be the future for these devices (even Corning knew that, and was abandoning their liquid crystal effort). LCs have many drawbacks, and MEMS can do everything LC can. The only question is when the future will begin for these components.
deer_in_the_light 12/5/2012 | 12:38:26 AM
re: Corning Chops Wavelength Blocker Products available on the market use free-space demultiplexing combined with an array of actuators based on liquid crystal or MEMS.
redface 12/5/2012 | 12:38:07 AM
re: Corning Chops Wavelength Blocker Thanks, jeb_knucklehead.

So I would say it is wrong that this article says individual channels can be dropped to near zero using the Corning device. It would be very difficult to control individual channels this way. The Corning product should be a band dynamic gain equalizer.
wiseguy 12/5/2012 | 12:38:03 AM
re: Corning Chops Wavelength Blocker Reading this press release, http://www.corning.com/photoni... you can see that the Wavelength Blocker individually leveled each channel of 50 GHz in either C or L band. It also had the ability to block any channel to 40 dB. The maximum through loss was no more than 6.0 dB. Those guys at Corning had a real winner. If the market isn't really there, then we all are in trouble....
Demander 12/5/2012 | 12:35:34 AM
re: Corning Chops Wavelength Blocker I believe the blocker is a version of the wavelength router switch marketed under the name "PurePath". Does this mean Corning is dumping Purepath as well?

Regarding MEMS vs. liquid crystal: yes, LCs have problems. So do MEMS. Early in the development of these switches folks started in LC. Corning went from a demonstrator to a product in about 12 months using LC and they have the market. I think that says something about LC. It's funny because when you mention LC technology everyone just goes into their drawbacks (slow speed, polarization sensitive) yet MEMS devices are often slower and have been much slower to market and have required larger development budgets. I think LC is a good technology but since it's old it just isn't as sexy. Problems with MEMS? (slow, lossy, difficult to control, low yield, stiction). Granted, these will (have been) solved but I think the jury is still out. MEMS is certainly favored for the giant cross connects, probably because polarization diversity is not required.

The basic form factor is to split the light using a grating and send each wavelength into a pixel on a linear SLM. The polarizations are split using birefringent prisms and handled separately (polarization diversity). Fiber 1's output is rendered vertical polarization and fiber 2's output is rendered horizontal and they are mixed on the pixel. If the LC is unswitched then 1 goes to 1. If it's switched 1 goes to 2 and visa versa. This all happens on a per-wavelength basis. It's cool.

JDS is also using the same basic approach. The original design patent comes from Tellium, who evidently went MEMS cross connect anyhow. Lzrjk is right, both Chorum and Spectraswitch are not wavelength selective to my knowledge. Anyone know how they are doing?
Sign In