[c-nsp] [External] Re: Support for CFP2

[Nick Hilliard]

Rob Evans via cisco-nsp wrote on 24/01/2024 23:27:
> Yeah, as I mentioned, there may be alternatives.  Noting that the OP wanted
> a range of 800km+, do SO also offer a suitable pluggable for the
> line-side?  The ones I could see from a cursory glance appear to be
> dispersion limited to 450km at 50GHz, or need 100GHz.

oh duh, I missed the 800km+ requirement bit. This is definitely the sort 
of area where for ease of implementation / longer-term support, getting 
an off-the-shelf 100G transport device would be useful. A DCI / 
transponder solution here would leave local hand-off to short-haul 
optics (lr4 / sr4 / aoc), which abstracts all the complexity / 
transceiver cost / etc away from any L3 device.

The OP would also need to figure out what's happening with regen in the 
middle. Is this OEO or optical-only amplification? Link characterisation 
becomes a thing once you're outside short-haul / metro connectivity.

Obviously this isn't to say that you can't do long haul on alien waves - 
you certainly can. But I wouldn't like to get involved when something 
goes wrong and everyone starts finger-pointing at everyone else about 
whose kit is acting the maggot.  Strategically it's usually simpler to 
abstract off potentially complex areas like this into their own 
self-contained box which can be managed as a discrete unit. Once you 
factor in the cost of managing complexity, it's usually no more 
expensive and often less.  All the more so in cases where the A and B 
ends of a link are different organisations.

There are plenty of market options for this. An internet search for 
"long haul dci" will give manufacturer names.

Nick