x
DWDM

Infinera Races the Clock on 100G

Infinera Corp. (Nasdaq: INFN) might have 100 Gbit/s ready a little sooner than expected.

During an earnings call Thursday, company officials said a product supporting 100Gbit/s wavelengths could get announced in the December quarter of 2011, with volume shipments coming by June 2012. Previously, Infinera had said only that 100 Gbit/s would come in 2012.

That new product will be based on Infinera's 500Gbit/s photonic integrated circuit (PIC), which carries five channels that run at 100Gbit/s each.

Infinera's 40Gbit/s product, which won't be based on its PIC, will ship in the September quarter, officials said Thursday. As analyst George Notter of Jefferies & Company Inc. pointed out, that's a bit later than the "mid-2011" target Infinera was using before. "Clearly I was hoping to get it out closer to Q2," CEO Tom Fallon told him on the earnings call.

Why this matters
Infinera needs that 100Gbit/s product. Analysts still believe the company's hyper-integrated optics can be a big difference-maker in telecom, but Infinera is late to 40 Gbit/s. Officials admit their sales have suffered -- either through lost business or customers waiting for 100Gbit/s -- and Infinera has lowered its revenue forecasts for three quarters in a row.

The company still has a chance to catch most customers just as they're ready for 100 Gbit/s, but its margin of error has gotten slim. Every quarter, it's important for Infinera to reassure customers and investors that it's going to deliver 100 Gbit/s on time.

For more
The latest on Infinera:

— Craig Matsumoto, West Coast Editor, Light Reading

Page 1 / 2   >   >>
Pete Baldwin 12/5/2012 | 5:06:21 PM
re: Infinera Races the Clock on 100G

CEO Tom Fallon made an interesting point about 40G at the end of the earnings call. A lot of the 40G opportunity is in China and therefore closed to Infinera (he didn't specify why, but obviously it's because of Huawei and maybe ZTE).  Another chunk of 40G is happening at AT&T and Verizon, two customers that happen to not use Infinera.


So, whatever percentage of the 40G market that adds up to -- that's a part of the market that Infinera had no chance at anyway.


Infinera's still losing business because of 40G, they admit as much on every conference call. But one analyst was asking how much available market would be added with a 40G product, and Infinera claims it's not as much as you'd think.


I still think they had to give up on 40G mainly because they just couldn't get it done in time, but ... if they can pull off 100G in a big way, maybe all will be forgiven.

ninjaturtle 12/5/2012 | 5:06:19 PM
re: Infinera Races the Clock on 100G



I am sure you'd agree that 40G is not the end game for carriers but a line to line need until an economical 100G solution is available. Based on the PIC integration of INFN's 100G product details (200+components squeezed into a chip) they seem to be the economical models needed to make the switch. In addition through ASIC technology their software and hardware will be a smarter more configurable product. Everyone else is using the same commercially available components on their 100G DWDM. Higher carrier cost and lower margins for the Box suppliers. PIC integration will prove its worthiness more incrementally with 100G then it ever did with 10G. Plus they are hitting the market at the very early adoption phase this time. Two huge advantages for INFN.



 

Pete Baldwin 12/5/2012 | 5:06:17 PM
re: Infinera Races the Clock on 100G

Ninja - Yep, I agree.


But for a company whose sales are volatile to begin with (and therefore eyed suspiciously by Wall Street), the business lost in the 40G generation certainly isn't a good thing.


They still might kick butt in 100G, but I doubt that skipping 40G was part of Plan A.

paolo.franzoi 12/5/2012 | 5:06:10 PM
re: Infinera Races the Clock on 100G

"CEO Tom Fallon made an interesting point about 40G at the end of the earnings call. A lot of the 40G opportunity is in China and therefore closed to Infinera (he didn't specify why, but obviously it's because of Huawei and maybe ZTE).  Another chunk of 40G is happening at AT&T and Verizon, two customers that happen to not use Infinera."



Stevery's comment made me re-read the messages.  I took a look at this and said to myself, is it not true that one can replace 40G with 100G in that passage?


seven


 

Stevery 12/5/2012 | 5:06:10 PM
re: Infinera Races the Clock on 100G

> They still might kick butt in 100G, but I doubt that skipping 40G was part of Plan A.


So remind me:  Is their 100G actually a 10G line rate or 25G line rate?  (And why are we naming DWDM according to the aggregate bandwidth again?)


Certainly there's a bunch more risk at 25G line rate if that's what they're planning.  Since they're already running 10G, stringing 10 of those together should be trivial, but they would get beat by the 25G guys.


They already got hammered at 40G line rate, which clearly requires a talent set that they do not have, so I'm curious which way their technical lead is steering.

Stevery 12/5/2012 | 5:06:05 PM
re: Infinera Races the Clock on 100G

Thanks Craig and IJD.


So really this is a 20x25G system.  The optics and electronics are all running at 25G, which is hopefully within the reach of Infinera, which has product experience running at only 10G.


I note that there were very few companies that pulled off an honest-to-god 40G line rate in the past decade (and Nortel was not one of them.)

Pete Baldwin 12/5/2012 | 5:06:05 PM
re: Infinera Races the Clock on 100G

Yes, it's dual carrier, with each carrier carrying two 25Gbit/s data streams:


http://www.youtube.com/watch?v...


Stevery does have a point -- we're going along with each vendor's aggregate bandwidth lingo, counting the throughput that they target per channel. We could stand to call them out on the dual-carrier approach more often.

IJD 12/5/2012 | 5:06:05 PM
re: Infinera Races the Clock on 100G

Based on what little information has been revealed, Infinera's 500G PIC solution seems to be based on 10 50Gb/s channels (using overclocked coherent 40G optical hardware?) not 5 100Gb/s channels as stated in the report -- in other words the same dual-carrier approach to 100G as Nortel/Ciena used.

IJD 12/5/2012 | 5:06:03 PM
re: Infinera Races the Clock on 100G

It depends what you mean by "real honest-to-God 40G line rate"...


Nortels' coherent 40G system uses ~12Gbaud optics, with 4 bits per baud DP-QPSK this needs ~20GHz optical channel bandwidth -- but it's certainly really transmitting 40Gb/s on the line as one signal without channel multiplexing, and was the "hero pioneer" for commercial coherent optical systems.


This gives far better tolerance to fiber imperfections than a non-coherent 40G solution using binary modulation with 40Gbaud on the fiber in a 50GHz channel, if this is what you mean by "real..." i.e. the solution the industry has largely walked way from, because it's great in the lab but not so good in real life.

Mark Sebastyn 12/5/2012 | 5:06:00 PM
re: Infinera Races the Clock on 100G

The only thing that matters is $ per bit per km per Hz. How much does it cost to send a quantity of bits (typically 100G or 10G) a given distance over a certain amount of optical spectrum.


Certain applications will favor certain technologies. Fiber type, distance, whether it is used as an overlay to non-coherent fibers or in greenfield all affect performance. There is no clear answer. Having a number of solutions in the quiver is really the "best" solution. I think Infinera as well as others are working in this direction.


I believe the INFN solution is a sincle carrier dual polarization approach, which is what most people are doing for 100G. Nortel/Ciena currently has a dual carrier dual polarization link but I'm sure they are working on the next step.

Page 1 / 2   >   >>
HOME
Sign In
SEARCH
CLOSE
MORE
CLOSE