The MPLS WG Archive

["Thomas D. Nadeau" <tnadeau@cisco.com>]

         Comments are inline.

>I have some doubts. Can you please clarify me.
>
>#mplsTunnelHopEntry  OBJECT-TYPE
>#SYNTAX        MplsTunnelHopEntry
>#MAX-ACCESS    not-accessible
>#STATUS        current
>#DESCRIPTION
>#"An entry in this table represents a tunnel hop.  An entry is created by a
>#network administrator for signaled ERLSP set up by an MPLS signaling
>#protocol."
>#INDEX { mplsTunnelHopListIndex, mplsTunnelHopIndex }
>#::= { mplsTunnelHopTable 1 }
>
>It has been proposed/suggested as above, the index definition for
>TunnelHopEntry in the forthcoming TE MIB.
>
>(Probably the MIB might be help understand better. Since it is in the
>draft stage and we use the indices in our discussion, I request you this 
>clarification.
>Sorry if I have missed some mails where you might have clarified, if so
>just please refer them).
>
>1) My understanding of the mplsTunnelHopListIndex, is that it is an index
>    to a list of ER Hops.
>    i.e. when the rows in the Tunnel Hop Table are to be created, I can
>    create rows with the following indices combination.
>    a) mplsTunnelHopListIndex = 1, mplsTunnelHopIndex  = 1
>    b) mplsTunnelHopListIndex = 1, mplsTunnelHopIndex  = 2
>    c) mplsTunnelHopListIndex = 1, mplsTunnelHopIndex  = 3
>    d) mplsTunnelHopListIndex = 2, mplsTunnelHopIndex  = 1
>    e) mplsTunnelHopListIndex = 2, mplsTunnelHopIndex  = 2
>    f) mplsTunnelHopListIndex = 2, mplsTunnelHopIndex  = 3
>
>   Now I have two ER Hop Lists which has 3 ER hops each.
>
>   Is my understanding correct?

         Yes, I believe that you are correct. However,
we have modified the tunnelEntry to contain an extra
object, which would be the primary index into the
mplsTunnelHopTable. This would obiate the need to
extend the indexing of the hop entries in the
same way in which the tunnelTable was extended recently.
This change has the added advantage of the tunnelHopTable
indexing being totally independent of how the tunnelTable
is indexed.

>2) If my above understanding is wrong, then the following questions
>    are not valid.
>
>    a) If I am right, then my questions are how are we associating the
>    ER Hops to the tunnels?

         We always have associated them.


>3) Though with the earlier MIB we were forced to configure the
>    ER Hop information again and again for different tunnels (when
>    a ER hop is shared by a set of tunnels, or otherwise set of
>    tunnels pass through the same ER Hop), It was straight forward
>    and easy to implement.  Clarifications will help me appreciate
>    the new hop table proposal.

         Exactly. If we use the indexing scheme described above,
it is very easy for multiple tunnels to point at the same set
of ERHops.

         --Tom