[c-nsp] NCS5501(-SE) in P and SP Peering roles

[adamv0025 at netconsultings.com]

> Jason Lixfeld
> Sent: Wednesday, November 15, 2017 12:48 PM
>
> The NCS5001 is out of scope here, as I was recently told that it’s not
> recommended as a P; it’s not (any longer?) a supported use case.
> 
No doubt, it's a L2 switch, which happened to do some MPLSy stuff.

> The NCS5501 as a P (ISIS, BFD, LDP) and the NCS5501-SE as a SP Peering (ISIS,
> BFD, LDP, v4/v6-full-Internet-table-in-a-VRF, Full XR RPL stack, CoPP/LPTS)
> would be the candidate use cases.
> 
Actually NCS5501 with 18Mpps per 10GE and NCS5501-SE with just 15Mpps per 10GE (slower because of the external TCAM) are both at the bottom end of my performance chart. 
So don't expect line-rate with all the edge protecting filters (or god forbid QOS in addition to that) -basically like MPC7 -but that card in my eyes is an MPLS core card.

So yes for P-core nodes the performance may be just right, but for edge I would not recommend it. 

Oh and there's the pre-classifier thing I'm not sure about on forwarding NPUs used in these, so if you plan on using it in a converged backbone (voice/data/internet) you might want to ask your SE what the situation is there (whether it can protect high priority data if ASIC is overloaded during DDoS which with such underpowered chip is very likely). 
 
adam