[TICTOC] Comments to draft-mirsky-mpls-residence-time-02

[Gregory Mirsky <[e-mail]>]

Dear All,
greatly appreciate comments we received after presenting our work to TICTOC and MPLS WGs.
Below my recollection of comments with responses:

*         Shahram pointed that TTL Expiration mechanism would alter residence timing comparing with regular forwarding of an MPLS packet that may be carrying time synchronization packet.
True. If RTM used over particular LSP, then all time synchronization packets may be carried over RTM ACH and LER may ignore accumulated residence in certain cases.

*         Yaakov and George expressed concern by possible performance impact that use of TTL Expiration method by RTM may have on installed LSRs that punt packets upon TTL expiration.
If an LSR does not advertise RTM capability in IGP TE, then it would not appear on RTM Set Object and, as result, would not be considered as "next downstream RTM LSR" by RTM capable LSR that calculates distance. If nevertheless non-RTM capable LSR receives RTM ACH packet, then it will drop it as the LSR would not recognize RTM ACH value.

*         Yaakov pointed that RTM capability is per port, not per node.
Yes, and that how we plan to advertise RTM capability in IGP TE.

We greatly appreciate all suggestions and comments.

                Regards,
                                Greg