italobusi
commented
5 months ago I think that the answer to this question depends on what are the N ports and this depends on the resiliency architecture.
Let's say that X is the number of router ports required to support unprotected IP traffic.
If the resiliency strategy is to protect only at the IP/MPLS layer, the number of router ports required is (X+1)<=Y<=2*N, depending on the percentage of IP traffic that needs to be protected.
In this case, the (N+1) router port protection would require Y+1 router ports and I would agree with you about the cost/benefit issues.
An alternative resiliency strategy can leverage the multi-layer aspects of POI and rely on optical network protection/restoration to recover most of the failure scenarios and the (N+1) router port protection to recover the failures of the links between the IP routers and the ROADMs.
In this case the (N+1) router port protection would require only X+1 router ports with clear cost benefits. If, as it happens in many cases, the probability of a failure on the links between the IP routers and the edge ROADMs is quite low compared with the probability of a failure within the optical network, the availability provided by this configuration can also be higher (at a lower cost) that a single-layer protection and also be able to recover 100% of the IP traffic
My 2 cents
Italo
As discussed in our call last Tuesday, cost vs benefits of this use case of having a spare port in the router and a spare transponder in the optical node just to be used in case of a router port failure is not so clear to me. Most common approach is to solve this at IP/MPLS layer (FRR, etc.). However it was agreed not to remove this use case without Fabio's agreement.
_Originally posted by @jbouqui153 in https://github.com/italobusi/draft-poidt-teas-actn-poi-assurance/pull/10#discussion_r1509159236_