& cplSiteName &

JDSU: MEMS Aren't Just Memories

Light Reading
OFC/NFOEC News Analysis
Light Reading
2/25/2002
50%
50%

JDS Uniphase Inc. (Nasdaq: JDSU; Toronto: JDU) turned heads today by announcing a "breakthrough" optical component that combines DWDM and multiport switching capabilities (see JDSU Unveils Optical Wavelength Switch).

The company says it hopes to build products based on this prototype that will cut the costs of making and deploying optical networking. The switch will reduce the space and cost of a DWDM-based switching fabric, JDSU says, while upping the amount of capacity available. Unfortunately, JDSU can't say when it will be able to do this. It isn't even ready with samples. "We cannot provide a schedule in terms of the progress and availability of the product," says a spokesperson.

Intriguingly, JDSU is basing its innovation not on spiffy, new science but on tried-and-true (and some say passé) MEMS technology, wherein thousands of tiny tilting mirrors direct light to its destination.

In recent months, industry sources have argued that MEMS (micro-electro-mechanical system) is old-fashioned due to its reliance on mechanical elements (see OMM and Altamar Buys a Bargain). Vendors who already offer MEMS-based optical switch modules, including OMM Inc., haven't shown signs of rapid market acceleration.

OMM says that, while its wares are being tested by many customers, it knows of just three trials underway of gear based on its devices. And Lucent Technologies Inc. (NYSE: LU), which makes the switch module for its own LambdaRouter optical switch, has released information on just two customers (see Lucent's LambdaRouter Turns Japanese) -- one of them, troubled Global Crossing Ltd. (NYSE: GX).

JDSU says it's got a new take on MEMS that allows it not only to switch light but to pick the wavelengths going from one port to another. This capability isn't yet featured in competing products such as those from OMM.

But OMM says it could beat JDSU to the punch. "They have yet to demo a MEMS-based optical switching module that meets Telcordia specs," says Conrad Burke, senior VP of marketing and sales at OMM. In contrast, OMM is already shipping a 32-by-32-wavelength switch approved by Telcordia Technologies Inc., which may make it easier to gain approval for new modifications.

Burke also questions whether JDSU will have any-to-any switching capabilities in a timely fashion. JDSU's news release says its prototype packs ten ports, switching 65 wavelengths in the L-band (1570 nm to 1620 nm) into a gadget that fits on a standard circuit card (JDSU won't give specific dimensions).

In contrast, OMM's 200-by-280-by-35mm switch can forward light to any of the 32 wavelengths (1290 nm to 1610 nm) in its fabric. Of course, the wavelength selection capability is yet to come, Burke acknowledges.

JDSU says future plans include higher port counts. Also, the new switch will work in the more conventional C band wavelengths (1520 nm to 1570 nm).

Observers are left with many questions. Can JDSU make its MEMS-based switch into a real product? Will OMM and Lucent make good on their existing wares?

Some say that, whatever happens, it's naive to dismiss JDSU's claims based on its use of MEMS. "MEMS is an essential technology that's proven and network-ready," says Saeid Aramideh, VP of product management at Iolon Inc., which makes tunable lasers. He says the market slowdown, not lack of viability, have hit makers of MEMS switch modules hard, noting that JDSU's news is "a step in the right direction... toward dynamic and flexible network architectures."

If anyone can make a viable product of a prototype like this, it's probably JDSU, which already has well established methods for making all kinds of optical components. JDSU also has testing tools, including its own swept laser system, which it says enabled it to make this prototype and would presumably allow it to develop real products effectively later on.

JDSU, Lucent, and OMM will all be exhibiting at the upcoming Optical Fiber Communication Conference and Exhibit (OFC)

— Mary Jander, Senior Editor, Light Reading
http://www.lightreading.com Want to learn more about the impact of testing and measurement on the manufacture of optical equipment? Tune in to Light Reading’s upcoming Webinar: "Optical Test Equipment in Manufacturing: Automation is Key," set for February 28. Click the following link to register: http://www.lightreading.com/webinars.asp#11506

(20)  | 
Comment  | 
Print  | 
Newest First  |  Oldest First  |  Threaded View        ADD A COMMENT
Page 1 / 2   >   >>
Multiplex
50%
50%
Multiplex,
User Rank: Light Beer
12/4/2012 | 10:53:17 PM
re: JDSU: MEMS Aren't Just Memories
Most MEMS stories -- including this one -- focus on the technical limitations or capabilities of MEMS. But where is the market?

The OXC market is overwhelmingly OEO right now, with no signs whatsoever of that changing. Most equipment vendors appear to be interested in ROADMs based on tunable technology. Small MEMS switches may make sense for some protection switching, but opto-mechanical and liquid crystal can give MEMS a run for its money there too.

Can the MEMS switching business really generate enough money for Agere, Corning, JDSU, OMM, Onix, C Speed, and Lord knows who else to flourish?
flanker
50%
50%
flanker,
User Rank: Light Beer
12/4/2012 | 10:53:12 PM
re: JDSU: MEMS Aren't Just Memories
OMM says that, while its wares are being tested by many customers, it knows of just three trials underway of gear based on its devices. And Lucent has released information on just two customers...

OOO simply isnt broadly deployed so you cant draw any conclusions about MEMS based on the limited penetration of ALL types of OOO switches.
This argument won't be settled for another four years.


optical_guy
50%
50%
optical_guy,
User Rank: Light Beer
12/4/2012 | 10:53:12 PM
re: JDSU: MEMS Aren't Just Memories
<jdsu a="" allows="" another.="" as="" but="" capability="" competing="" featured="" from="" going="" got="" in="" isn't="" it="" it's="" light="" mems="" new="" not="" omm.="" on="" one="" only="" pick="" port="" products="" says="" such="" switch="" take="" that="" the="" this="" those="" to="" wavelengths="" yet="">

Am I missing something here.....it sounds like they have merely integrated a mux/demux with the MEMS....oh, yipeee.

OG

</jdsu>
diffraction
50%
50%
diffraction,
User Rank: Light Beer
12/4/2012 | 10:53:08 PM
re: JDSU: MEMS Aren't Just Memories
Yes, there is something missing here

Having to splice in a mux/demux with a 8x8 or
32 x 32 port switch count sounds costly to me even if the fiber is ribbonized. What customers want is a mux/demux fully integrated with a wavelength switching fabric without any fiber routing issues. I'm sure that JDSU has integrated these functions in a low cost manner.
slayer666
50%
50%
slayer666,
User Rank: Light Beer
12/4/2012 | 10:53:07 PM
re: JDSU: MEMS Aren't Just Memories
The entire issue is infact that you can integrate MEMs devices with AWGs or other demux (echelle grating) waveguide solutions to create photonic "systems on chip" ASICs. It is not going to be a revolution however...it will be an evolution. Devices will be built as discrete elements, then integrated as process technology allows. This stuff is in its infancy, like a transistor in the 50s/60s...but it will come faster as better tools and concepts exist today.
Look at next generation platforms like SOI for the future of integrated Optical MEMs.
sdarch
50%
50%
sdarch,
User Rank: Light Beer
12/4/2012 | 10:53:06 PM
re: JDSU: MEMS Aren't Just Memories
>> This stuff is in its infancy, like a transistor in the 50s/60s...but it will come faster as better tools and concepts exist today. <<

agree. We'll wait and see when that will happen. For the news itself, just JDSU marketing guys doing their job.
diffraction
50%
50%
diffraction,
User Rank: Light Beer
12/4/2012 | 10:53:06 PM
re: JDSU: MEMS Aren't Just Memories

The real question I would ask is where is the huge market for all these 256x 256 or 32x32 large cross connects? and not where is the market for
MEMS.

Unlike OMM, JDSU doesn't sound like they are using MEMS for a large cross connect.

This is the way I think about it...

There is quite a difference between a train station with 32 individual train tracks each serving one train and a train station with a few tracks serving 64 trains. Which sounds more
realistic to you.

Any confused VCs out there ?
Tony Li
50%
50%
Tony Li,
User Rank: Light Beer
12/4/2012 | 10:53:05 PM
re: JDSU: MEMS Aren't Just Memories
The real question I would ask is where is the huge market for all these 256x 256 or 32x32 large cross connects? and not where is the market for MEMS.
--------------


The application for such cross connects is as a POP interconnect. You plug all of your DWDM, ADMs, routers, test gear, etc. into the OXC. And now, through a little bit of magic, you have a fully automated patch panel that allows you to reconfigure your network on a whim. This allows you to cut provisioning time. You can conceivably automate you network reconfigurations. You can do wavelength engineering, not just traffic engineering.

Seems incredibly useful to me.

Tony
Multiplex
50%
50%
Multiplex,
User Rank: Light Beer
12/4/2012 | 10:53:04 PM
re: JDSU: MEMS Aren't Just Memories
Tony:

I see your point, but is there any evidence that service providers are interested in buying anything like this?. In my experience there is resistance to automation of provisioning and, in any case and no real sense of how much money that would actually save. In addition an automated patch panel of the kind you describe would cost considerably more than regular patch panels. Not to mention that service providers are quite suspicious of all-optical switching technology.

In these days of limited CAPEX and low initial costs, is there really a viable market for this kind of product? I think the long-term (four or five years) answer is probably yes. My real question is whether those companies that are banking on this type of application now -- expecially the start ups -- can really afford to wait that long. I think for most of these companies the answer is no, although JDSU and Corning have the resources to wait.

For those who disagree with this assessment, the question is not to describe how useful this kind of product could be, but rather, considering all the factors that going into selling this kind of product, just how big the market for the kind of subsystems that JDSU is offering is going to be in 2002 and 2003.

RGreg
50%
50%
RGreg,
User Rank: Light Beer
12/4/2012 | 10:53:04 PM
re: JDSU: MEMS Aren't Just Memories
I see your point, but is there any evidence that service providers are interested in buying anything like this? In my experience there is resistance to automation of provisioning and, in any case and no real sense of how much money that would actually save.
-----------------------------------------

In a word, yes.

Most of the big service providers are laying out roadmaps now for all-optical (or at least mostly optical) networks. I've seen the powerpoints, and the OXCs population is high. Also, I've talked with at least one person whose job it was to determine how much that carrier would save in the migration steps (and he sounded very positive about it). So clearly they're planning it out.

I think the MEMS guys are looking to be the first to get the big contract. You're right in that it will take several years for this stuff to get implemented, and they'll take it in steps (although the RBOCs may take smaller steps than some other carriers). BUT, it's important to get their product into the planning now & not several years from now.


On another note, why did Lightreading just reference Lucent and OMM? Are they the only two companies still surviving in the MEMS space? I did a quick check of Lightreading's recent press releases and found Umachines and Megasense. Also, I seem to recall Onix landing a big funding round last fall...
Page 1 / 2   >   >>
Featured Video
Flash Poll
Upcoming Live Events
September 17-19, 2019, Dallas, Texas
October 1-2, 2019, New Orleans, Louisiana
October 10, 2019, New York, New York
October 22, 2019, Los Angeles, CA
November 5, 2019, London, England
November 7, 2019, London, UK
December 3, 2019, New York, New York
December 3-5, 2019, Vienna, Austria
March 16-18, 2020, Embassy Suites, Denver, Colorado
May 18-20, 2020, Irving Convention Center, Dallas, TX
All Upcoming Live Events