The MPLS WG Archive[Date Prev][Date Next][Thread Prev][Thread Next] [Date Index][Thread Index][Author Index][Subject Index] LSR MIB comments and questions
-----Original Message-----
>From: Thomas D. Nadeau [mailto:tnadeau@cisco.com]
>Sent: Wednesday, April 05, 2000 3:23 PM
>To: Cucchiara, Joan
>Cc: 'mpls@uu.net'
>Subject: RE: LSR MIB comments and questions
>
> Hi,
>
> Sorry about the delay in response. It was a
> long flight back from Adelaide! ;)
> > >I also question why this extra variable is required to
> > >convey the label space membership attributes.
> > >
> > >If mplsInterfaceConfIndex is zero, then it uneqivocally
> > >has the attributes 'in global label space' and 'not in
> > >local label space'.
> > >
> > >If mplsInterfaceConfIndex is non-zero, that it uneqivocally
> > >has the attribute 'in local label space' and the 'in
> > >global label space' variable arbitrates that (optional)
> > >attribute.
> >
> > Yes, but there are cases in ATM where an interface needs
> > to be part of both label spaces. This is why we have
> > provided two separate variables here.
>
> The combination of the index and ONE of these
> objects would (I think) allow all combinations. For example,
>
> Yes, that is true, although it implicitly notes
> which interfaces are per-interface and which are per-platform.
> Our intention was to be explicit about this. The reason is
> that if you use the implicit version, it is unclear as to
> how the implementation should note the label ranges. We do
Tom,
Could I suggest combining these two boolean objects into
one enum object, mplsInterfaceLabelParticipationType,
with values ( perPlatformOnly, perInterfaceOnly, both)??
DESCRIPTION:
"If the value of the mplsInterfaceConfIndex for this
entry is zero, then this object MUST be perPlatformOnly.
If the value of the mplsInterfaceConfIndex for this entry
is non-zero, then this object may be either 'perInterfaceOnly'
or 'both'. If the value is perInterfaceOnly then this means that
the value of the MIN/MAX IN/OUT Label Range objects for
this entry contain valid values. If the value is both, then
the value of the MIN/MAX IN/OUT label Range objects do not contain
valid values (???? OR maybe need to contain the same values as
the MIN/MAX IN/OUT Label Range objects for the
entry which has the mplsInterfaceConfIndex of zero????)"
[NOTE: The description needs work because it needs to be
co-ordinated with the MIN/MAX IN/OUT Label objects.]
I think this enum would be explicit enough (and also combines
the two boolean objects into one enum object -- so one less object!).
What do you think?
Thanks,
Joan
> agree however, that changing the names to IsPerPlatformLabelSpace
> and IsPerIfLabelSpace is a great idea as it makes this consistent
> with the architecture.
>
> --Tom
|
|