x
Mergers & acquisitions

Nortel Shuts Optical Switch Effort

Nortel Networks Corp. (NYSE/Toronto: NT) has ceased work on the monster all-optical crossconnect it acquired with Xros Communications in March 2000.

"Nortel has decided not to bring the OpteraConnect PX to market," says spokesman David Chamberlin. Nortel had planned to start commercial trials of the switch later this year.

Chamberlin says "dramatic changes in market conditions" caused Nortel to finally throw in its monogrammed Xros towel. Carriers just aren't interested in big optical switches right now (see No Riches From Optical Switches ).

Sources say Nortel had additional reasons to pull the plug on Xros. It's been an open industry secret that the product's R&D was stymied. And given ongoing delays in its most strategic new products, Nortel is challenged to focus its efforts on more mundane OEO solutions (see Nortel's HDX: The Future Under Fire ).

Chamberlin says it's not yet clear what effect closing Xros will have on Nortel's financials, but he notes that part of the company's $12.4 billion June 2001 writedown was attributable to goodwill associated with the acquisition (see Nortel Buys: Reaping the Whirlwind?).

An undisclosed number of employees who worked for Xros have been laid off. Greg Reznick, the startup's CEO, who has been serving as "president of the Xros division" since the buyout, will continue to report to Brian McFadden, Nortel's president of optical long-haul networks, in an as yet undefined role. "What his next role will be is unclear," Chamberlin says (see When the CEO Hits the Road).

Nortel was criticized for paying too much when it bought Xros for about $3.25 billion in stock, a transaction that amounted to paying about $36 million for each of the startup's 90 employees. Nortel scoffed, citing the deal's basis in stock, not cash, and insisting the price was fair (see Nortel Buys a Monster Crossconnect).

Originally, Nortel hoped to release a massive 1000x1000-port all-optical switch via its Xros deal. But as the months dragged on, it became apparent that the technology had some problems. Sources say the 3D MEMS (micro-electro-mechanical system) components on which Xros originally based its design were unique, but impractical to implement in anything except a massive long-haul switch -- exactly the opposite of what the market wants today.

Despite the canning of the Xros project, sources inside and outside the company say Nortel plans to make the best of what it bought. Efforts are underway to merge elements of the Xros crossconnect with smaller switch chips, they maintain. The results, of course, remain to be seen. Meantime, the waters have closed over the once-bright promise of Nortel's "jawbreaking" all-optical switch.

— Mary Jander, Senior Editor, and Stephen Saunders, Founding Editor, Light Reading
http://www.lightreading.com
<<   <   Page 9 / 10   >   >>
trixie 12/4/2012 | 10:49:30 PM
re: Nortel Shuts Optical Switch Effort So, if 5 dB is not low enough insertion loss for ULH all-optical, what is an acceptable number?

______________________________________


Well, let's start with the VSR capabilities, which is around 3dB or so. More than one 5dB node in a link and you're pretty much forced to do 3R. 3-4 hops at 2-3 dB/hop is reasonable.

so <3dB would be for openers. I know of one technology right now that is <2 dB, but it will be in need of a good control plane to be useful.

As our religious Zealot, Perry, indicates- managment (automation, restoration, QoS) is key.
trixie 12/4/2012 | 10:49:30 PM
re: Nortel Shuts Optical Switch Effort Perry1961 orated:

This advertisement was brought to you by a true believer,LOL!

__________________________________________

Yes, no doubt a heavy investor, too based on your "fundamentalist" posturing :-)

edgecore 12/4/2012 | 10:49:30 PM
re: Nortel Shuts Optical Switch Effort Thanks for the answer, wish I could digest half of it.

These questions may be lame, but here they go:

But not having any regens, wouldn't you need to pop those OOO switches down in the middle of a very powerful long haul system?

Does Tellium provide LH systems? They must have transponders or something?

Does Tellium have any lasers on their All Optical boxes or just waveguides?

If you only have OOO, how the hell do you drop traffic? Or do yu always drop a full OC48 or OC192 to an awaiting big Sonet Box that in turm demuxs it all for you...

I must be missing the point here, again, please guide me...

EC
trixie 12/4/2012 | 10:49:29 PM
re: Nortel Shuts Optical Switch Effort Straight shooter-

All excellent points. The control planes and the traffic planes will become separare entities, each on its own developmnent curve. Even OEO and electronic circuits will be managed out of band, eventually, IMHO.

I suspect the hybrid configuration will win out, as you rarely will need all grooming or all photonic switching at any given node.

The trick re performance monitoring, is *how*. I agree with the premise- these are really analog transmission systems that just happen to carry digital signals- the electronics are either good, or fried, and I suspect the optical plant introduces the majority of signal defects (notice there is no proclamation of expert authority here) Root cause measurement (optical domian) makes sense. I imagine the EMS SW could handle the issue of flaky electronics if necessary.

As has always been the case, we are just waiting on technology to catch up to the ideas. Then the trick is to pick the right technology to deploy.
trixie 12/4/2012 | 10:49:28 PM
re: Nortel Shuts Optical Switch Effort Edgecore asks:

If you only have OOO, how the hell do you drop traffic? Or do yu always drop a full OC48 or OC192 to an awaiting big Sonet Box that in turm demuxs it all for you...

_________________________________________

Correct- the difference being that it does not have to go through similar sonet boxes at each node, enroute to its final destination.

Main driver here is cost- line cards (15xxnm regenerators) cost about $50k each, street price. each wavelength in the DWDM system would require 2 of these cards per node. You do the math. you also save on all the filter mux/demux cards at each node, not cheap, either.

Yes, some equipment providers have developed a single card solution. You still need one card, per wavelength, per fiber degree of the node.

TELM offers only a switch. they have two different products, one is OEO, with OC-48 grooming, and 512x512 ports. The other is a smaller OOO device, around 32x32 ports. They may have a bigger unit, using Corning MEMs out, but I really don't pay that close attention to the press releases.

Others, who offer both switching and transport, like SCMR, CIEN, Innovance, NT (someday), are starting to integrate colored optics (DWDM line cards) directly on the switch chassis. This saves an additional interface between the DWDM and OXC, usually a 1310nm SR serial interface.

Getting very low loss OOO switches will be valuable, as you can then do multiple hops across more standard long-haul systems. Most MEMs systems are around 5-7 dB, although I believe IMMI recently announced they would have something in the 3dB range.

I hope that doesn't make it worse....
Maxwell loggerhead 12/4/2012 | 10:49:27 PM
re: Nortel Shuts Optical Switch Effort "But the IMMI 3 dB does not include any closed loop control optics, which adds a couple of dB at least."
______________________________________________

You are incorrect, Sir. The IMMI switch has closed loop control optics and it does not add "a couple of dB." In addition, The IL is less than 3dB.

-Max
self 12/4/2012 | 10:49:27 PM
re: Nortel Shuts Optical Switch Effort But the IMMI 3 dB does not include any closed loop control optics, which adds a couple of dB at least.
trixie 12/4/2012 | 10:49:26 PM
re: Nortel Shuts Optical Switch Effort Maxwell jumps in defense:

You are incorrect, Sir. The IMMI switch has closed loop control optics and it does not add "a couple of dB." In addition, The IL is less than 3dB.
___________________________________________

Mea Culpa,

the IMMI web page cites an IL for their new 80x80 switch as 1.7db typical, with a max of 3.0. This is without connectors*, so add another 0.3dB per connector pair. So depending on your arguements of rIL measurement techniques, total loss is 2.0-3.6dB.

I think a variance of almost 2X in IL is a bit loose, and perhaps a bit unrealistic. If the max is 3dB, then 3dB is what I would design to in the network. The rest is marketing and semantics.

Maxwell, care to tell us if it is still the garden variety 3D MEMs?

How does it pass zone 4 compliance for shock and vibration?

Maxwell loggerhead 12/4/2012 | 10:49:25 PM
re: Nortel Shuts Optical Switch Effort The webpage need to be updated. If you want to know more: OFC booth 1647. We will have our shock tester there dropping our mirror assembly, so you can see first hand how they do vs shock. Its a good product. no smoke, all mirrors.
self 12/4/2012 | 10:49:24 PM
re: Nortel Shuts Optical Switch Effort "But the IMMI 3 dB does not include any closed loop control optics, which adds a couple of dB at least."
______________________________________________

You are incorrect, Sir. The IMMI switch has closed loop control optics and it does not add "a couple of dB." In addition, The IL is less than 3dB.

_____________________________________________

Wow, so the switch fabric does 0.7ish dB to 2.0ish dB? That is quite a box. What are you bringing to OFC? The 80 and/or the 256?
<<   <   Page 9 / 10   >   >>
HOME
Sign In
SEARCH
CLOSE
MORE
CLOSE