& cplSiteName &
Comments
Threaded  |  Newest First  |  Oldest First        ADD A COMMENT
PO
PO
12/4/2012 | 11:21:36 PM
re: Alcatel Releases Core DWDM Gear
"...when a (generally common) fiber gets cut, what happens to *all* the downstream channels??"

Can someone answer the same question for copper? How would a cut in a copper line be different than a cut in a fiber line?
---
Copper networks drop timeslots, or don't have amps in common. Fiber networks pass multiple wavelengths through the same amps on a common fiber.

My comment was referring to the instantaneous power spike on *other* channels at such common amps. At an extreme, let's say an amp (EDFA) is handling 128 channels: 127 from a 'thru' path, and one from a recent add/drop point. A fiber cut takes out the 127, but the one channel is unaffected.

At the amp, the input power drops by about 21dB (assuming equalized channel powers). Will the surviving channel's output power spike instantaneously (and if it does, how do we ensure it doesn't damage a receiver)? Will the amp think that all 128 channels are gone, and trip off? How does the single failure (fiber cut) not couple over to otherwise unaffected channels?

So I guess the questions are, how would this new gear eliminate failure mode coupling, and what failure mode coupling exists for copper?
dwdm2
dwdm2
12/4/2012 | 11:21:37 PM
re: Alcatel Releases Core DWDM Gear
"...when a (generally common) fiber gets cut, what happens to *all* the downstream channels??"

Can someone answer the same question for copper? How would a cut in a copper line be different than a cut in a fiber line?
OEOEOEOEOEO
OEOEOEOEOEO
12/4/2012 | 11:21:42 PM
re: Alcatel Releases Core DWDM Gear
I managed a NOC that used the Alcatel DWDM (1640). Beware! too many anomalies. I guess Alcatel finally gave up on their promise to provide Raman on the 1640!
st0
st0
12/4/2012 | 11:21:42 PM
re: Alcatel Releases Core DWDM Gear
Beware! too many anomalies. I guess Alcatel finally gave up on their promise to provide Raman on the 1640!
======
What kind of anomalies? thx.

-st
st0
st0
12/4/2012 | 11:21:43 PM
re: Alcatel Releases Core DWDM Gear
"My other concern is that 192 channels require some pretty powerful optical amplifiers, and some good link engineering so I really question if anyone will deploy that many with the OPEX making it impractical, but the selling point is that they could."
========
agree. possibly got EDFA plus extended EDFA and Raman (either distributed or discrete one... more likely to be the distributed Raman) amps to ensure the gain tilt capability is somewhat in control.

-st
PO
PO
12/4/2012 | 11:21:43 PM
re: Alcatel Releases Core DWDM Gear
I'm actually a bit more worried about the common equipment: what happens to a link when an amp fails (after all, with that many thru channels, we're talkin' about some serious optical power spikes.)

Or, when a (generally common) fiber gets cut, what happens to *all* the downstream channels??
dwdm2
dwdm2
12/4/2012 | 11:21:56 PM
re: Alcatel Releases Core DWDM Gear
"25GHz spacing will buy you up to ~160 channels in each band, as long as your filters and wavelockers are up to the task. So maybe they did extend C band a little?"

Actually I'll tend to think that it is hard to find a 25 GHz filter if you're talking about TFF or FBG. For AWGs though it is OK, but they sold their planar staff to AVNX. Anyone knows what kind of filter they are using? My guess is still the 50 GHz ones.
dwdm2
dwdm2
12/4/2012 | 11:21:56 PM
re: Alcatel Releases Core DWDM Gear
"So for the 1626 they had the option to extend below the bottom end of the C-Band or above the L-band by 15nm or do both, but say by 8nm at each end to fit in 192 wavelengths for 1.92 Tbps."

OK, I stand correct. However, it is difficult to envision that systems will run wavelengths in the C-L crossover region or below C and above L region. It is not primarily a question of a smart amplifying scheme or even low-loss link engineering. It is the fiber that has higher loss in those regions. So even if the channels are equalized at the wdm, transmission loss within the fiber is pretty non-uniform over all the channels, making it close to impractical (in addition to the OpEx).

So if they are just increasing the number of channels from 160 to 192 keeping everything else same (still a spectral efficiency of 0.2 by your count), it may as well be a marketing strategy.

If, OTOH, they are accommodating all 192 channels within the C band, by increasing the spectral efficiency to say 0.4, that is something important. Any thoughts on this?
chi2
chi2
12/4/2012 | 11:21:58 PM
re: Alcatel Releases Core DWDM Gear
Yes, they do mention G.709. "Optical Transport Network."

192 channels @ 10G each-
25GHz spacing will buy you up to ~160 channels in each band, as long as your filters and wavelockers are up to the task. So maybe they did extend C band a little?

192 is a strange number to use- it's either a marketing thing or software limited, or their rack just ran out of space to handle more channels. The announcement did make a big deal out of footprint.
MrLight
MrLight
12/4/2012 | 11:21:59 PM
re: Alcatel Releases Core DWDM Gear
In response to dwdm2 pot 2 in this thread:
"Because it is OC-1 (if I'm not wrong), they can fit it in a 25 GHz spacing. Actually for 0.25 nm spacing filters, one can get more than 31 GHz to play, and upto 200 channel over the C-band. At 52 Mbps, this is a spectral efficiency of 0.17.

So the whole thing is news worthy simply because it has all 192 channel in one box? Not sure if I missing something obvious. Can anybody clarify?"

I can clarify; you are definitely missing something.

The wavelengths can be up to 10Gbps each. With up to 192 of them across both the C-band and L-band.

Alcatel's previous record was 160 wavelengths (1.6 Tbps) in their 1640 DWDM product.

The 1640 had:
80 wavelengths @ 50 Ghz (.4nm spacing) for the C-band 1529 to 1565 nm (36nm)
80 wavelengths @ 50 GHz for the L-band - say 1567 to 1605nm (38nm) Note: You could run it to 1620nm).

Note: they would need to waste at least 1.6 to 3.2 nm between where the C-band ends to where the L-Band begins for the bandsplit filter.

So for the 1626 they had the option to extend below the bottom end of the C-Band or above the L-band by 15nm or do both, but say by 8nm at each end to fit in 192 wavelengths for 1.92 Tbps.

MrLight...shedding some light on the subject

P.S.One thing to take note is that I am not aware of any commercial analyzer that can display/measure all 192 channels at once.

My other concern is that 192 channels require some pretty powerful optical amplifiers, and some good link engineering so I really question if anyone will deploy that many with the OPEX making it impractical, but the selling point is that they could.

Another thing about this announcement is there is no mention of support for G.709 FEC which would have the line rate be 10.7 Gbps per channel. Nortel has FEC so Alcatel still has some catching up to do if they doing. If they do, than they should have put it in their press release.
pavlovsdog
pavlovsdog
12/4/2012 | 11:22:00 PM
re: Alcatel Releases Core DWDM Gear
Are all wavelengths in C-band at 25 GHz spacing? Or split between C- and L- bands?
dwdm2
dwdm2
12/4/2012 | 11:22:00 PM
re: Alcatel Releases Core DWDM Gear
It seems like they have 192 of OC-1 to make up the 9.98 Gbps. So speed wise it is a backward technology?

Because it is OC-1 (if I'm not wrong), they can fit it in a 25 GHz spacing. Actually for 0.25 nm spacing filters, one can get more than 31 GHz to play, and upto 200 channel over the C-band. At 52 Mbps, this is a spectral efficiency of 0.17.

So the whole thing is news worthy simply because it has all 192 channel in one box? Not sure if I missing something obvious. Can anybody clarify?

Thanks.


Featured Video
Upcoming Live Events
December 3-5, 2019, Vienna, Austria
December 3, 2019, New York, New York
March 16-18, 2020, Embassy Suites, Denver, Colorado
May 18-20, 2020, Irving Convention Center, Dallas, TX
All Upcoming Live Events