Re: Question on InetAddressType

["C. M. Heard" <heard@pobox.com>]

On Tue, 13 May 2003, Andy Bierman wrote:
> At 03:14 PM 5/13/2003, Wijnen, Bert (Bert) wrote:
> >In the MPLS MIB arena, a question is coming up on a
> >statement in a MODULE0-COMPLIANCE aka:
> >
> >    OBJECT mplsFTNAddrType
> >    SYNTAX       InetAddressType { ipv4(1), ipv6(2) }
> >    DESCRIPTION
> >        "An implementation is only required to support IPv4
> >         and IPv6 addresses."
> >
> >The question now is: what does that mean for a box that only
> >supports IPv4 and not (yet) IPv6. One could write multiple
> >compliance statements, but that seems a bit too much does it not.
> >
> >Do we consider this acceptable:
> >
> >    OBJECT mplsFTNAddrType
> >    SYNTAX       InetAddressType { ipv4(1), ipv6(2) }
> >    DESCRIPTION
> >        "An implementation is only required to support IPv4
> >         and/or IPv6 addresses. An implementation only needs to
> >         support the addresses it actually supports on the box."
> 
>                       ^^^^^^^^^ address types
> 
> I think this is acceptable.  The IETF should not mandate that
> devices support either IPv4 or IPv6.  The market will decide
> if the product supports the right address types.  Normally
> we see this subrange mandate ipv4 and omit ipv6.  This statement
> above allows for the possibility that IPv6 is supported, but
> not IPv4.

I agree with Andy.  Before I saw this I probably would have written
an OBJECT clause like this without a SYNTAX clause and made the
DESCRIPTIOn clause a little more explicit, but the version above is
clearer (and takes a lot less effort to write correctly).

Of course, if the object in question is read-only this is something
of a hair-splitting exercise since the implementation gets to choose
the values it returns, and I would not worry about fixing it in that
case.  But the suggestion above is a good one for read-write or
read-create objects, and for index objects in read-create tables I
like the technique of writing OBJECT clauses like this and stuffing
them inside the MODULE-COMPLIANCE DESCRIPTION clause, either plain
or formatted as ASN.1 comments, as was done in
draft-ietf-sigtran-sctp-mib-09.txt.

Mike