RE: Where to root MIB modules - descriptors

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

On Tue, 6 Jan 2004, Wijnen, Bert (Bert) wrote:
> Two things to think about:
> 
> 1. For the case of Mike's text below:
>    > >    b) vendor creates a proprietary "stable snapshot"
>    > 
>    > vendor allocates fizbinExp from its own tree
>    > 
>    > root OID is { fizbinExp 1 }
>    > module name is VEND-EXP-FIZBIN-MIB-SNAPSHOT-1
>    > 
>    I would prefer  to see them as OID is { vendFizbin 1 }
>    I.w. also in the module root OID make it clear it is from vendor "vend"

OK

>    ANd I assume you mean that "VEND" gets replaced by the name of the
>    actual vendor, no? Same for "vend"

Yes.

>    > descriptors are as in WG draft
>    >  
>    Doable... I'd rather see them prefixed with "vend"

This is really up to the vendor.  I've noticed that some indeed do
as you suggest.

> 2. For the allocation under experimental, we have experience with 
>    RFC2786. We can consider that RFC more or less as a 
>    "stable snapshot" can we not?

I think this document was rather more than just a stable pre-release
snapshot of a draft;  it was intended for production deployment.
Indeed, at one time the cable-modem industry had that MIB module as
"mandatory to implement" for cable modems that were to be certified.

>    ANd when we suggested they move it to stds track, they are 
>    VERY RELUCTANT (I think they are just refusing) to even reroot 
>    under mib-2. They want to stay under experimental tree!
> 
>    So what does that tell us?

I think it means that the experimental subtree should be reserved
for the purpose set forth in RFC 2578, i.e., "to identify objects
being designed by working groups of the IETF."  In other words, it
is just for stuff that is under development in IETF working groups.
Experimental OIDs would not be handed out for specifications such as
RFC 2786.  For those, the IETF would ask the vendor consortium that
sponsors the spec to get an enterprise number and register the MIB
module under it (this is what the ATM Forum has done).

Of course, there is nothing to prevent people from taking a WG
draft, implementing it, and then shipping it to customers ... but if
we make the policy very clear up front (OID has to move from
experimental to mgmt when module enter the standards track) they'll
have no right to complain.

//cmh