x
DWDM

All-Optical Smackdown in Secaucus

SECAUCUS, N.J. -- At an optical networking conference held here on the fringes of the Meadowlands, several switching players took the opportunity to pound the all-optical proponents into the cold turf.

The conclusion from a number of participants and audience members in a panel on optical switching held here at the CIROptics 2002 conference: Given the current levels of carrier spending, all-optical switching is years away.

The panel included companies building optical switches with electrical cores (OEO switches) -- Tellium Inc. (Nasdaq: TELM), Ciena Corp. (Nasdaq: CIEN), and Altamar Networks Inc.; and two startups touting an all-optical approach (OOO) -- switch vendor Calient Networks Inc. and MEMS startup Transparent Networks Inc..

The bigger guys promptly pummeled the little guys.

Krishna Bala, CTO and cofounder of Tellium, pointed out several hurdles to the all-optical approach. The first problem has to do with manufacturability. “Cost and integration are cheaper for OEO switches today,” said Bala. “OEO ASICs reuse benefits of CMOS manufacturing. The process is already set up, so we can leverage the manufacturing.”

In a rare moment of solidarity, Tellium and Ciena, both currently shipping OEO switches, said that all-optical switches won’t be widely deployed in major carrier networks for three to five years. Altamar concurred.

Mark Lutkowitz, vice president of optical networking research with Communications Industry Researchers Inc. (CIR) and chair of the panel, said he doesn’t expect them to ship in volume for five to seven years. Lutkowitz said that for certain applications, like ultra-long-haul express forwarding, all-optical switching can be cost effective, but in widescale applications it just doesn’t make economic sense.

“On paper, all-optical is a slam dunk,” said Lutkowitz. “As for ultra-long-haul deployments, that is only about 15 percent of the total market. Corvis and Calient are obsessed with all-optical. [Corvis CEO, David] Huber is in denial.”



Corvis Corp. (Nasdaq: CORV), which has already been shipping its all-optical switches to Broadwing Inc. (NYSE: BRW), did not attend the discussion (see David Huber Pipes Up).

Lonely Calient, the only all-optical switch vendor sitting on the panel, tried unwaveringly to convince the audience that all-optical switches are just around the corner.

“To say that all-optical switching won’t be here for another five years is a fallacy,” said Tim Dixon, vice president of marketing for Calient. “I think we’ll see volumes kick up in 2002.”

Dixon argued that the technology helps carriers save on operational costs to such an extent that they cannot afford to ignore wavelength switching. He told the audience that his company would actually be shipping and deploying its product into an unnamed carrier’s network by midyear.

Gabriel Lowy, vice president of technology research at Crédit Lyonnais Securities Inc., said the whole lot may be trompé. After the discussion, Lowy pointed out that the big problem with optical switching in general, whether it be OEO or OOO, is that it doesn’t create new revenue streams for carriers. Lowy, whose bank lends money to carriers, said that service providers are most interested in creating new ways to make money. He felt the discussion missed that point entirely.

“Carriers spent all this money building out networks, and now they need to show shareholders how they are going to justify the expense of the past several years,” he said in an interview after the discussion. “Yes, reducing cost is important, but finding new revenues is their priority. I didn’t hear that here.”

At any rate, the slowdown in capital spending does not spell out a particularly bright future for all-optical switching, at least in the near term. The real answer of when all-optical becomes viable may not be apparent until carriers return to significant capital spending on new networks.

— Marguerite Reardon, Senior Editor, Light Reading
http://www.lightreading.com
Page 1 / 5   >   >>
wdog 12/4/2012 | 11:04:23 PM
re: All-Optical Smackdown in Secaucus I believe Mr. Lowy is viewing the world a little too black and white. Optical switches, particularly STS-1 grooming switches are the foundation for many carrier services. Grooming switches provide carriers a platform which will support rapid provisioning of leased line services from T3 through OC-192. (Lease lines are still the bulk of revenue generating services for many carriers).The ability to provision these services rapidly cuts carrier costs and allows them to offer these services in a more timely manner, AND under shorter term contracts. Adding features like GMPLS based software recover will support different types of SLAs. Faster response to demand, shorter term contracts, flexible SLAs, and over time bandwidth-on-demand are all sources of new revenue for carriers made possible by optical switches.
The_Holy_Grail 12/4/2012 | 11:04:22 PM
re: All-Optical Smackdown in Secaucus The article really needs to emphasize the bottom line foundation for deploying optical switches. If there isn't a economic incentive for deploying them, customers won't go in and replace/upgrade their existing equipment.

Also, its not clear in the article on how one would manage optical swicthes since the demarc points often cross boundaries and then requires customized SLAs.

Yes, I think there is a need for optical switches but also believe it will be tightly coupled to ASIC solutions.
Kozmoz 12/4/2012 | 11:04:22 PM
re: All-Optical Smackdown in Secaucus Marguerite Reardon (Following the Herd??)Hmmmmmm......I always thought that unless you're the "LEAD DOG".......The view is always the same.

At least Huber and the Corvis Team aren't staring up the "brown eye" of played out technology.

flanker 12/4/2012 | 11:04:20 PM
re: All-Optical Smackdown in Secaucus "In a rare moment of solidarity, Tellium and Ciena, both currently shipping OEO switches, said that all-optical switches wonGÇÖt be widely deployed in major carrier networks for three to five years."

They are just repeating what the carriers are telling them.

Yes, reducing cost is important, but finding new revenues is their priority. I didnGÇÖt hear that here.GÇ¥

Moot point. Carriers are slashing capex across the board. Worldcom said in the Salomon Smith Barney conference in Scottsdale last week they were effectively cutting capex in half, from 9.5 bln in 2000 to 5.5 bln in 2002 in order to boost cash flow.











straight shooter 12/4/2012 | 11:04:19 PM
re: All-Optical Smackdown in Secaucus I attended this session early Monday morning in beautiful Secaucus, NJ and this is not how it went at all?!? Were we in the same session?

In fact, browsing the badges at the check in table during the session, I noticed that both this article's author and Scott R. had yet to pick up their attendance badges(?).

Were you really there? Or are you just parroting the "objective" research of Lutkowitz (NOT).
hawkman 12/4/2012 | 11:04:18 PM
re: All-Optical Smackdown in Secaucus They are kicking everyones "tail" in the OEO switch market. OEO switches are big, expensive and not to mention the 40 or 50K for the line cards. The carriers want to make more money, OOO offers them the opportunity to save money on Capital expenditures. This translates to lost revenue for the incumbent equipment makers. If you dont have to convert signals to Electricity at each city, you dont need all that equipment and carriers are happy because they will save hundreds of millions of dollars. The carriers need to hear this and put pressure on the system vendors. OOO is coming, adapt or die.
Litewave 12/4/2012 | 11:04:18 PM
re: All-Optical Smackdown in Secaucus Author: hawkman
the line cards. The carriers want to make more money, OOO offers them the opportunity to save money on Capital expenditures. This translates to lost revenue for the incumbent...


On the surface it may appear that an OO based solution provides capex savings, in reality though, the oustanding issues with a pure OO based solution currently negates any capex savings versus an OEO based solution.

If you don't know that already, then you I'm sure you must be desperately pulling out your hair trying to figure out why OO based solutions (eg. Lucent's Lambda Router) aren't selling! FACT, not fluff.
SiO2 12/4/2012 | 11:04:17 PM
re: All-Optical Smackdown in Secaucus Litewave writes:

> On the surface it may appear that
> an OO based solution provides capex
> savings, in reality though, the
> oustanding issues with a pure OO
> based solution currently negates
> any capex savings versus an OEO
> based solution.
>
> If you don't know that already,
> then you I'm sure you must be
> desperately pulling out your
> hair trying to figure out why
> OO based solutions (eg. Lucent's
> Lambda Router) aren't selling!
> FACT, not fluff.

I certainly wouldn't extrapolate
the lack of revenue generated by
the (ironically named) Lambda Router
to infer any failure in the value
proposition of OOO switches.

As has been explained in other
threads, the combination of an
OEO device (to switch local traffic
and groom subrate traffic) and an
OOO device (to switch/route express
traffic) is the most cost effective
approach to building a regional or
long-haul network. By leveraging
the transparency of OOO and the
service agility of GMPLS, providers
are able to future-proof their
deployments while reducing OPEX
in the near term. QED.

SiO2

Physical_Layer 12/4/2012 | 11:04:15 PM
re: All-Optical Smackdown in Secaucus Y2K, nice post. I completely agree with the two statements you made below:

"On average, 70% of the total traffic that comes to a switch node is meant for the adjacent node and only 30% is meant for local consumption"

"The experts are correct. There is no future for ALL-optical but there is also no future for ALL-electrical"


y2k 12/4/2012 | 11:04:15 PM
re: All-Optical Smackdown in Secaucus The problem with all-optical is that it is ALL-optical.

A photonic switch does not mean ALL-optical. A photonic switch is a device that allows the connection of two fibers without having to first change the photonic signal into electrical signal. A photonic switch is not functional equivalent to an electronic switch but a photonic switch is a lot cheaper than an electronic switch.

There is no reason to build an ALL-optical network because there is no way to extract revenues out of a consumer without first converting a photon into an electron. But the network can be a lot cheaper to the end users if there are as few photon-electron conversions as possible except in the case of revenue extraction.


All data traffic is long distance.

On average, 70% of the total traffic that comes to a switch node is meant for the adjacent node and only 30% is meant for local consumption. Most of the transit traffic is wavelength-level and most (if not all) of the local traffic is sub-wavelength. Wavelength-level traffic is suitable for low-cost photonic switching while sub-wavelength-level traffic requires electronic switching.

If the network makes no distinction between transit and local traffic, then the comsumers have unknowingly incurred unnecessary expenses related to photon-electron conversions that have nothing to do with revenue extraction. Studies have shown that this hidden cost penalty is as much as 50% of the total cost.

The future belongs to a hybrid architecture where photonic and electronic switches work in concert (what Ciena calls All-Optimal and what Corvis calls Multi-Band).

The experts are correct. There is no future for ALL-optical but there is also no future for ALL-electrical.


--y2k--
Page 1 / 5   >   >>
HOME
Sign In
SEARCH
CLOSE
MORE
CLOSE