RE: MIB module numbering for TBD MIB roots

["Andrew Smith" <ah_smith@pacbell.net>]

How about some lateral thinking here: just make the problem go away by being
sensible about allocating OIDs *during* the proposed-standardisation process
instead of at the end of it.

1. Enterprise assignments are of no concern to us here - they are for each
enterprise to manage and it's not IETF's job to stop implementors shooting
themselves in the foot with their own gun.

2. As soon as a draft MIB is adopted by an IETF WG as a work item, assign it
a root under mib-2 at some appropriate level. If a WG-adopted MIB is later
dropped then either give back the OID to the free pool or else deprecate it
in IANA's list (these are not really a scarce resource).

The current situation where you have to wait for IESG approval of the draft
as an RFC is quite unnecessary: it still does not stop the not-infrequent
cases where an implementor ships a product prematurely and does not bother
to check what MIBs crept in; it does not handle versioning (i.e. changes to
a previously approved MIB) which is an issue that SMIng needs to tackle.

But I'm probably tilting at windmills here ... I've brought up all these
points before.

Andrew Smith


-----Original Message-----
From: owner-mibs@ops.ietf.org [mailto:owner-mibs@ops.ietf.org]On Behalf
Of Andy Bierman
Sent: Thursday, January 17, 2002 11:03 AM
To: Juergen Schoenwaelder
Cc: rmonmib@ietf.org; mibs@ops.ietf.org
Subject: Re: MIB module numbering for TBD MIB roots


At 10:26 AM 1/17/2002, Juergen Schoenwaelder wrote:

>>>>>> Andy Bierman writes:
>
>Andy> I like plan (d):
>
>Andy> Create an OID rooting convention for Standard and Enterprise
>Andy> MIBs, requiring MIBs-in-progress to be rooted under a branch
>Andy> called 'unassigned'.
>
>Andy> For MIBs in Internet Drafts (mandatory):
>
>Andy>    unassigned ::= { mib-2 0 } fooMib ::= { unassigned x } ( == {
>Andy> mib-2 0 x } to keep IMPORTS stable)
>
>Andy> For vendor MIBs (suggested):
>
>Andy>    <prefix>Unassigned ::= { <enterprise-id> 0 } fooMib ::= {
>Andy> <prefix>Unassigned 0 x }
>
>Andy> e.g.  ciscoUnassigned ::= { enterprises 9 0 } cFooMib ::= {
>Andy> ciscoUnassigned x }
>
>Andy> where 'x' is a temporary assignment, which may be reused over
>Andy> time
>
>Requires even more edits to get rid of the unassigned definitions.

I think it's more important that the numerous MIB readers
(and even MIB tools) identify a MIB root as unassigned, than it is
to save the 1 MIB writer 10 seconds of editing.

>I still think { mib-2 0 } or { mib-2 xxx } are good enough - it is
>virtually impossible to prevent someone from doing the wrong thing.
>I do not think (d) is in any way stronger than (a) or (b) - perhaps I
>mis-understand the meaning of 'x'.


The 'x' is the leaf integer of the MIB root assignment.
This is needed to allow more than one unassigned MIB to exist
in a development environment, since multiple MIB modules using
the same root won't work in most (all?) SNMP tool-kits.


>/js

Andy