>I would like to ask a fundamental question here. >Do we need transit counters for only MPLS-SR, or do we need it for MPLS-LDP as >well? Would be good to start with SR-MPLS label stack, including transit nodes – not all LSRs but the Nodal/Adj-Sids as indicated in the SR-Label stack.
>Also WRT the proposal to do ingress collection, if nodal paths are used, that >only tells us the approximate path, not the hotspot which I understand to be >the original goal. 1. Though it would be limited – SR path level stats would be extremely useful – especially slicing use cases where 100’s of thousands of flows gets mapped relatively less number of SR-Paths/Sr-NODAL-ADJ-SID stack. 2. If we are fixated on the “hotspot issue” solution – then you *may* get into a more complex solution (both w.r.t RLD & MLDs and with usage of special labels) … #2 is not new (possible with existing LDP network). But #1 would be a great plus for ST-TE deployments. -- Uma C.
_______________________________________________ spring mailing list [email protected] https://www.ietf.org/mailman/listinfo/spring
