Re: [Tsvwg] Re: [IETFMIBS] TCP Extended Statistics MIB, and v6ops TCP/UDP Process MIB

[Matt Mathis <[e-mail]>]

If there is not a technical problem with two tables that AUGMENT the same base
table, I strongly advocate advancing the documents separately.   ESTATS mib is
almost ready for the RFC editor queue.  -mib-issue- is not even a WG item yet,
and may or may not converge quickly.   I really don't want to reopen ESTATS
mib for lots of new revisions at this time.

That said, the issues raised in -mib-issue- are real.  We were aware of them
but (mistakenly?) believed that there did not exist a standardizable OS
independent solution.  I would like to see this work proceed, but not at the
cost of stalling ESTATS.

Thanks,
--MM--

On Fri, 12 Jan 2007, Kristine Adamson wrote:

> Juergen,
>    Sorry for the confusion.  The current draft of the TCP-ESTATS-MIB is
> at:
> http://www.ietf.org/internet-drafts/draft-ietf-tsvwg-tcp-mib-extension-14.txt
>
> The TCP/UDP Process MIB draft is at:
> http://tools.ietf.org/wg/v6ops/draft-persson-v6ops-mib-issue-01.txt
>
> Kristine Adamson
> IBM z/OS Communications Server: TCP/IP Development
> Phone: [phone]   T/L 444-7911
> Internet e-mail:[e-mail]
>
>
> Juergen Schoenwaelder <[e-mail]> wrote on 01/12/2007
> 08:13:50 AM:
>
> > On Fri, Jan 12, 2007 at 06:07:43AM -0700, Kristine Adamson wrote:
> > > I've worked with Matt Mathis a bit on the tsvwg group's TCP Extended
> > > Statistics MIB because I work on an IBM product's SNMP support.  As
> far as
> > > I am aware, you may have more than one "augmenting" table which
> augments a
> > > base table.  But an augmenting table can not itself be a base table. I
> am
> > > copying the MIB mailing list for any additional feedback.
> > >
> > > So, both the tcpEStatsConnectIdTable (from the tsvwg's TCP-ESTATS-MIB)
> and
> > > the tcpConnectionInfoTable (from the v6ops' TCP-PROC-MIB) can augment
> the
> > > tcpConnectionTable in the TCP-MIB.
> > >
> > > That said, maybe the TCP Creation MIB objects in the TCP-PROC-MIB's
> > > tcpConnectionInfoTable, could be moved to one of the tables in the
> > > TCP-ESTATS-MIB.  We will be looking into implementing the
> TCP-ESTATS-MIB
> > > in the future in order to support additional information for TCP
> > > connections.  Since the TCP-ESTATS-MIB is in the final approval
> stages,
> > > maybe these MIB objects could be added to one of the TCP-ESTATS-MIB's
> > > table in the future, as an updated version of the MIB module.   Or,
> maybe
> > > they should be added to the TCP-MIB's  tcpConnectionTable?
> > >
> > > Unfortunately, I don't think any IETF group is working on an extended
> > > statistics MIB module for UDP endpoints.  If we create a MIB table
> with a
> > > general name like, udpEndpointInfoTable, then you would expect that
> any
> > > other additional management data for UDP endpoints would be added to
> this
> > > table.  Instead, maybe it would be better to either add the UDP
> Creation
> > > MIB objects to the existing udpEndpointTable in the UDP-MIB, or rename
> the
> > > udpEndpointInfoTable to something more specific, like
> > > udpEndpointCreationTable.
> >
> > It would help people like me who did not follow any of this work if
> > you can provide us the names of the documents in question.
> >
> > /js
> >
> > --
> > Juergen Schoenwaelder       {International|Jacobs} University Bremen
> > <http://www.eecs.iu-bremen.de/>    P.O. Box 750 561, 28725 Bremen,
> Germany
>