The MPLS WG Archive

[francis arts vh624 8492 <fart@sh.bel.alcatel.be>]

Hello,

While reading the draft-nadeau-mpls-vpn-mib-00.txt, I had the questions
listed below. Could somebody provide an answer to these questions?

Thanks for your answer.

Kind regards,

    Francis.

--------------

        ************
        *Questions *
        ************

1) Section 8.

Section 8 of draft-nadeau-mpls-vpn-mib-00.txt states the following:
*) The text refers to the applicability of the Interfaces group to MPLS.
This section indicates that the mplsVpnInterface is stacked on an mpls
interface. On its turn the mpls interface is stacked on an 'underlying
layer'.
**) Later on in the document the text (of the mplsVpnInterfaceConfIndex)
indicates that there is only a one-to-one relation between the
mplsVpnInterfaceConfIndex and the row for the mpls interface in the
ifTable for MPLS / BGP VPN enabled interfaces.

However, considering the case of a CE - PE interface without the
carrier's carrier feature, the VPN traffic is transported as IP packets
over this interface. Therefore, it should also be possible to stack an
mpls VPN enabled interface on top of the underlying layer (PPP,
ethernet, ...) without an mpls interface in between?



2) mplsVpnInterfaceConfTable.

The name of this table suggests the notion of a VPN. So if a specific
interface is a member of 2 VPNs, this would imply 2 rows in the
mplsVpnInterfaceConfTable.

However, is the notion of a VRF rather than a VPN applicable here? [The
consequence would be that for an interface that is a member of 2 VPNs,
only 1 row is created in the mplsVpnInterfaceConfTable.]




3) mplsVpnVrfConTable.

A specific mplsVpnInterfaceConfIndex cannot belong to 2 rows in the
mplsVpnVrfConTable?




4) mplsVpnVrfConTable.

The text of the 'mplsVrfName' object indicates that 'this is the human
readable name of this VPN'.

However, should the term 'VPN' rather be 'VRF'?