RE: registration OIDs: do the values matter?

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

On Sun, 23 Feb 2003, Wijnen, Bert (Bert) wrote:
> [On Sun, 23 Feb 2003 Andy Bierman wrote:]
> > At 11:19 AM 2/23/2003 +0100, Wijnen, Bert (Bert) wrote:
> > >> >Indeed.  And I see a couple of assignments in that list that are
> > >> >wrong and need to be corrected.  Specifically,
> > >> >
> > >> >SSPM-MIB:     sspmMIB               M-I     { mib-2 777 }
> > >> >RAQMON-MIB:   raqmon                M-I     { mib-2 6889 }
> > >> >
> > >> >are listed in the corresponding I-Ds as
> > >> >
> > >> >SPM-MIB:      sspmMIB               M-I     { rmon 777 }
> > >> >RAQMON-MIB:   raqmon                M-I     { rmon 6889 }
> > >> 
> > >> these MIB modules are under development and will be fixed soon.
> > >> 
> > >They should NEVER have had these numbers in. [Instead] they should do:
> > >
> > >> >SPM-MIB:      sspmMIB M-I  { rmon xxx } -- to be assigned by Someone
> > >> >RAQMON-MIB:   raqmon  M-I  { rmon xxx } -- to be assigned by Someone
> > 
> > there's a practical problem with this approach -- this is illegal SMI.
> > MIB compilers don't like it. That's why we use bogus numbers instead.
> > I usually use 999999 to make it clear the number is bogus.  Other
> > MIB authors don't seem to do this.
> > 
> Not good either. It may be a tidbit more obvious ([to] MIB experts)
> that a 99999 is probably just a placeholder. But the proper way is
> to use something like xxxx or nnn or such.
> Anyone who wants to compile and implement better need to go
> [through] some manual process so it is VERY CLEAR that they are
> taking risks and better use non-existing OIDs.

For compilation I do what Andy suggests, but I try to make sure that
I do it in private -- in other words I try to make sure that it does
not appear in an I-D or even in an e-mail to someone else (I have
slipped on at least one occasion in e-mail during a MIB review).

If there is a reason to use a real OID for a document that is still
a draft, then it might be a good idea to acquire an OID in the
experimental branch.  We don't do that very much any more (and in
one case it was mistakenly done for something that was intended to
be deployed), but I think that is what the experimental branch is
for.  I agree that OID assignments in the mgmt branch should wait
until IESG approval.

//cmh