Analyst: AlcaLu's 100G Game-Changer

Alcatel-Lucent (NYSE: ALU)'s announcement that it has a commercial 100-Gbit/s optical product ready to ship today has moved the optical equipment market on in a number of ways, and will further squeeze the timeframe in which 40-Gbit/s systems will be deployed, according to Heavy Reading senior analyst Sterling Perrin. (See AlcaLu Goes Commercial With 100G and 40G: Time for the Third-Party Candidate to Bow Out?)

"That this is a commercial release, and not just talk about a product roadmap, is quite a surprise. Alcatel-Lucent must have stepped up its development cycle, and that's significant," says the analyst.

"It also means there are now two choices for carriers. Nortel [now part of Ciena Corp. (NYSE: CIEN)] has so far been able to say it's the only company with commercial 100 Gbit/s, and of course it has a customer in Verizon Communications Inc. (NYSE: VZ). Now it has Alcatel-Lucent for company." (See Ciena/Nortel Product Plans Revealed and Nortel Launches 100G Solution.)

The other significant thing about the launch, says Perrin, is that AlcaLu is "doing this with a single carrier, a single laser, whereas Nortel's gear has two sub carriers [two lasers] that generate two wavelengths, but which looks like one 100-Gbit/s wavelength.

"Now, the Nortel system still meets the requirements of fitting into the 50GHz spectrum, and there's no indication at all that this doesn't perform as well as a single carrier system. But major carriers, such as AT&T Inc. (NYSE: T) and Verizon, have been referring to single carrier in terms of trials, and they'll likely prefer to go down that route," the analyst continues.

"This gives Alcatel-Lucent something of an advantage just now in the market. This is going to put pressure on Ciena to deliver a single-carrier version of its 100G offering."

So who else might join Ciena and AlcaLu soon? Huawei Technologies Co. Ltd. , says Perrin, "has previously been quite slow, but recently it has been talking about 2011 availability, which is an accelerated timeframe. That gives Alcatel-Lucent a time advantage, but I expect today's news will probably speed Huawei up even more."

The analyst also expects to see developments from NEC Corp. (Tokyo: 6701), which has been talking about 100G for some time, and Nokia Networks , as well as Infinera Corp. (Nasdaq: INFN). (See Infinera Ditches 40G, Talks 100G.)

And then there's Cisco Systems Inc. (Nasdaq: CSCO). (See Cisco Renews Optical Focus With CoreOptics.)

The industry consensus seems to be that meaningful deployments and product revenues will begin in late 2011 and early 2012 -– Ciena reiterated that timeframe again today during its earnings conference call –- but the sales cycles are long for such heavy-duty infrastructure platforms, so AlcaLu might well be hitting the optical high street at just the right time. (See Ciena's Q2 Let Down by MEN.)

Perrin believes AlcaLu's timing "is good. At the recent Light Reading Packet Optical conference in New York, 100 Gbit/s was the big, and sometimes the only, talking point." (See LR Touts POTE Success, Scenes From Packet-Optical Evolution, Part II, Scenes From Packet-Optical Evolution, Part I, Google: 100 Gbit/s? How 'Bout 8 TBs Per Second?, Packet-Optical Evolves, and Verizon: Give Us More Flexible ROADMs for 100G.)

Perrin adds: "There's a real consensus in the industry that it's important to get commercial, Coherent 100G products to the market quickly."

— Ray Le Maistre, International Managing Editor, Light Reading

Page 1 / 3   >   >>
Stevery 12/5/2012 | 4:32:53 PM
re: Analyst: AlcaLu's 100G Game-Changer

Nortel's gear has two sub carriers [two lasers] that generate two wavelengths, but which looks like one 100-Gbit/s wavelength.

Funny.  Most physicists would say that a signal with two wavelengths looks like a signal with, in fact, two wavelengths.

Either those Nortel guys are teh awesome, or perhaps counting is done differently by marketing.

Stevery 12/5/2012 | 4:32:49 PM
re: Analyst: AlcaLu's 100G Game-Changer

So, in this case, it doesn't really matter what physicists would think but it does matter what operators think.

Don't fall for the snowjob:  Operators care more than just about the spectral efficiency, there's a lot of care and feeding for every wavelength added.  Two lasers means double the monitoring.  2x the alarms, and 2x the number of false alarms.

Which is a shame, cuz I was just on the verge of pitching Drew on my 100 packed lambda transceiver.  "Really Drew, it's just one wavelength.  It's revolutionary! Operators are gonna love it.  How about $20M?"

paolo.franzoi 12/5/2012 | 4:32:49 PM
re: Analyst: AlcaLu's 100G Game-Changer


Just a question, somebody made a post or article recently about 100G and how that the ITU grid was going to change.  One question to ask would be is 2 sub-carriers going to fit in the new scheme. 



LR_viewer 12/5/2012 | 4:32:49 PM
re: Analyst: AlcaLu's 100G Game-Changer

Compared to the 1-wavelength 100G approach, the 2-wavelength 100G approach doubles the number of key components required for coherent detection: (1) 2 polarization-diversity optical hybrids; (2) 8 ADCs to sample the I/Q components of each of the two polarizations of each of the two wavelengths; and (3) two independent sets of ASIC for DSP. This is in addition to the doubled number of lasers as well discussed here.


So, there are obvious cost/size/power advantages of the 1-wavelength (or single-carrier) approach. One key reason that single-carrier 100G was not previously commercialized is the lack of a single ADC-DSP ASIC capable of sampling and processing an entire 100G channel. Apparently, this technology challenge has been addressed by Bell-Labs empowered Alcatel-Lucent.

Performance wise, 1-wavelength 100G seems to offer higher nonlinear tolerance to neighboring 10G OOK channels to better support the migration from 10G to 100G.  

Sterling Perrin 12/5/2012 | 4:32:49 PM
re: Analyst: AlcaLu's 100G Game-Changer

So, in this case, it doesn't really matter what physicists would think but it does matter what operators think. And I don't have the answer to that question yet, but it should become clear pretty quickly now.

The NT approach uses 2 lasers but one 50 GHz channel, so it adds capacity to the fiber - unlike inverse muxing approaches of 10x10G. I agree with Ciena/NT that this distinction is a big one. It's 2 lasers, one channel.

What I don't know is which approach provides better performance and which costs less. I think these are the important questions.


Sterling Perrin 12/5/2012 | 4:32:48 PM
re: Analyst: AlcaLu's 100G Game-Changer

LR viewer,

Thanks for the detail on this.

I realize that cost to build and selling prices are 2 different things but I would have expected the ALU 100G to sell for much less than the NT 100G - based on fewer components. It's kind of reading between the lines at this point, but it *seems* like ALU'a pricing will be at least a little higher than NT's price and certainly not significantly less.

On card density, I need to take a closer look at the two. A bit of an aside, but the OpNext "card" that was used in the AT&T 100G trial from March literally took up an entire bay. It's no small achievement to get this into the 3 slots that ALU has announced. And I'm wondering how much the state of OpNext's 100G developments influenced Cisco's decision to buy CoreOptics.


Sterling Perrin 12/5/2012 | 4:32:48 PM
re: Analyst: AlcaLu's 100G Game-Changer


I have vague memory of reading that somewhere but can't find it - maybe it related to speeds beyond 100G? For 100G, a big piece of the value proposition is to work with the existing network, which is built around 10G specs and 50GHz spacing. For 400 Gbit/s or 1 Tbit/s maybe there's an opportunity to make some changes, but in telecom the "clean slate" approach never seems to be a viable option, since everything's built on some kind of legacy.



LR_viewer 12/5/2012 | 4:32:47 PM
re: Analyst: AlcaLu's 100G Game-Changer <div>Sterling,</div>
<div>You made many good points. I agree with you that price is not necessarily determined by cost alone. My guess is that ALU's 100G price will be more determined by the willingness-to-pay of&nbsp;the customers when factors such as performance (on both the transponder module level and the system level), time to the market, &nbsp;and future scalability are all taken into consideration.</div>
flake 12/5/2012 | 4:32:46 PM
re: Analyst: AlcaLu's 100G Game-Changer


Another important question is&nbsp;what is the dispersion-limited performance of&nbsp;the ALU 100G product.&nbsp;&nbsp;Have they integrated eDC techniques for&nbsp;compensating standard G.652 link&nbsp;GVD as well as improved DGD tolerances?&nbsp; I would assume so if the costs are thought to be higher than the Nortel 100G product.&nbsp; Any insights???



opticaljunkie 12/5/2012 | 4:32:46 PM
re: Analyst: AlcaLu's 100G Game-Changer

NT's 2 carrier approach uses 10g electronics &amp; optics which is cheap. ALU's single carrier 100g will need really high bandwidth components. That's usually the most expensive stuff. In terms of density, if ALU's single carrier is really 3 slots, NT's approach is still better since it's only 2 slot.&nbsp;

Opnext trial was conducted with fpga arrays, they don't have a real time receiver asic.&nbsp;

Page 1 / 3   >   >>
Sign In