DWDM on 250 Km dark fiber without re-amplification

Brandon Martin lists.nanog at monmotha.net
Sun Dec 25 08:41:12 UTC 2016

On 12/23/2016 07:14 PM, Jeremy wrote:
> Hi all,
> First, i'm sorry for my english, i'm french and i don't have a good
> level in this language. But i want some informations and i'm sure,
> someone will be give the good anwser about my question.
> So, i'm regarding to rent a dual dark fiber in France, the estimated
> distance is 225 Km, but i know there are a lot of optical switching on
> the highway where it's fiber is installed (in theory, all 80 Km). So, i
> used the bad scenario, in adding 25 Km on my need.
> I would like to buy a amplificator and multiplexer DWDM to add some
> 10Gb/s waves on this dark fiber. I've see that the amplification is
> better on 100 Gb/s synchronised ports, but we don't have enoug capacity
> on our router to add 100 Gb/s interfaces.
> So, someone has installed this type of hardware on a dark fiber without
> regeneration  on 250 Km of distance ?
> If yes, with what kind of hardware ? If you are commercial for this
> hardware, please contact me in private message.

Look up Raman amplification.  The short of what this does is it pumps a 
ton of power into the near end of the fiber span and creates what looks 
somewhat like a typical color-blind amplifier somewhere several dozen km 
out on the span.  You'll also need to dump a ton of power into the span 
at the far end using an EDFA or similar.  Even with both of those, that 
distance is still going to push the raw optical power budget of even 
most state-of-the-art transceivers especially if the fiber is old or of 
low quality (high loss, high dispersion, etc.).

The longest span I've ever gotten a vendor to commit to an engineered 
design for was about 140km, and of course they needed full 
characterization of the span before they'd do it.  At those distances, 
distance alone is no longer sufficient to throw together a design.

It seems highly likely that there's at least one re-gen facility along 
that span.  I'd definitely see if there is one and if you can get some 
space in it.  That will knock you down into the 100-130km range on both 
sides of the re-gen, hopefully, which is perfectly doable.

You are somewhat correct that 100Gb interfaces often handle longer 
distances better, but it's because they are often using coherent 
receivers and carrier-synchronous transmitters rather than raw power 
receivers and ASK pulsed transmitters.  There are vendors that sell 
coherent 10Gb transceivers, too, and they'll be cheaper than 100Gb 
solutions especially if you don't need the extra capacity anyway.  I'd 
definitely check them out for this type of application especially if you 
can't get any dispersion compensation in the middle since coherent 
optics are usually much more tolerant of chromatic dispersion.

The big vendor I've worked with in the past on this sort of stuff is 
Ciena (and they're certainly a juggernaut in the industry) though I have 
no connection to them other than as a satisfied (if occasionally broke 
after a PO or out of breath after seeing a quotation) customer/integrator.

Brandon Martin

More information about the NANOG mailing list