Re: MIB Reviewers

[Andy Bierman <abierman@cisco.com>]

At 02:17 PM 10/17/2002 -0400, Harrington, David wrote:

>Hi, 
>
>Is anybody reviewing the bridge mib drafts? 
>
>I don't know whether I should volunteer or recuse myself, since my boss is part of the IEEE 802.1 committee, and the mib is being edited by an Enterasys employee. 
>
>I just heard that between -01 and -02 revisions of the RSTP mib drafts, they added an object and changed the mappings between OIDs and descriptors. That normally wouldn't be too bad in an I-D, but the IEEE RSTP functionality has been approved for a year, so some vendors are already implementing the mib based on the I-Ds (despite the warnings).
>
>Is it considered acceptable by this group to change OID assignments between I-D revisions? 

Yes.

We just had a long debate on this subject on the Cisco MIB Police
mailing list.  IMO, work-in-progress means not-done-yet.  Vendors
are warned in boilerplate text in every I-D that they should not
rely on the stability of the I-D. We have an official procedure
at Cisco for snapshoting such I-Ds (change MIB root, module name 
and all descriptors in an algorithmic manner).  WGs should not have to 
follow SMI change rules until the MIB is released in an RFC.

I realize that this is tough on vendors, and made worse because
the IETF is so slow, but that's too bad. Any vendor that implements
a work-in-progress assuming it won't change is taking a big chance.

>Thanks, 
>dbh 

Andy