RE: MIB EDU training in Montreal

["Orly Nicklass" <orly_n@rad.com>]

Hi,

Since the audience might be people who write protocols but have no
intension to write MIB in the future I would extend issue 2 "when and
why" to the following:
I do believe that manageability is something that need to be thought out
even before MIB is written. As such it will be a good practice to plan
it in the protocol draft itself.

You can take as an example the PWE work where management is addressed at
drafts before the MIBs were written.
If people will address the concept in the protocol related drafts it
will ease the work on the MIB writers.

Keywords such as configurable parameters or monitored parameters can be
conceptually addressed in such drafts.

I think it will be wise to recommend such practice while discussing that
section.

Orly

-----Original Message-----
From: owner-mreview@ops.ietf.org [mailto:owner-mreview@ops.ietf.org] On
Behalf Of Wijnen, Bert (Bert)
Sent: Thursday, May 18, 2006 04:31
To: David T. Perkins
Cc: mreview@ops.ietf.org
Subject: RE: MIB EDU training in Montreal

Thanks David.
DO you have any slides/materials for the additional material you want
covered/extended?

Bert

> -----Original Message-----
> From: David T. Perkins [mailto:dperkins@dsperkins.com]
> Sent: Wednesday, May 17, 2006 19:46
> To: Wijnen, Bert (Bert)
> Cc: j.schoenwaelder@iu-bremen.de; mreview@ops.ietf.org
> Subject: RE: MIB EDU training in Montreal
> 
> 
> HI,
> 
> First, I may not be present on Sunday of the IETF meeting.
> If I was definitely going to be present, I would commit to help out.
> 
> On the document...
> There are many separate isssue that are covered by the document and 
> include:
>  1) a little background on SNMP and management info
>  2) guidance for when and why development of MIB documents
>     should occur in a WG
>  3) guidance on the scope of the content of a MIB module
>  4) the process and tools for writting MIB modules and the
>     I-Ds that contain them
> 
> Here are my suggestions:
> 1) I'd include just a little more background on SNMP and
>    management information. This is needed because of the
>    constraints of the protocol have a major affect on the
>    design of the MIB module definitions. Also, since there
>    are many different approaches to management, each with
>    different terminologies, it would be good to define
>    the key terms to help reduce confusion.
> 2) On the "when and why", the key points are:
>    a) Until you have deployment experience, it is difficult
>       to predict with great accuracy what management
>       information is needed to efficiently manage the technology.
>       Thus, the WG should try to get early prototype developments
>       implemented to provide feedback.
>    b) What management information is needed may affect the
>       protocol development, and thus, definition of the
>       management information must start before the protcol
>       work is completed. However, there is no need to
>       start definition of management information until
>       work on the managed protocol is well under way.
> 3) On the scope, this goes back to time/cost tradeoffs.
>    Monitoring of status and statistics is a must, and
>    providing notifications to reduce the latency and
>    to increase the scaling are desirable. Actions and
>    configuration is typically a large increase in
>    time and development. 
> 4) I hope that we can strongly encourage use of XML and
>    a standard XML template for the MIB document.
> 
> Hope this helps.
> 
> Regards,
> /david t. perkins
> 
> 
> 
> On Wed, 17 May 2006, Wijnen, Bert (Bert) wrote:
> > Thanks for the input from Juergen, Dan and DbH.
> > Other pls chime in if possible. 
> > I am willing to do an update of the slides that takes input into 
> > consideration. But I need that input rather sooner than later, 
> > becuase the 3 weeks prior to the IETF meeting I will be on vacation.
> > 
> > Bert
> > > -----Original Message-----
> > > From: Juergen Schoenwaelder [mailto:j.schoenwaelder@iu-bremen.de]
> > > Sent: Wednesday, May 17, 2006 08:13
> > > To: Wijnen, Bert (Bert)
> > > Cc: Romascanu, Dan (Dan); mreview@ops.ietf.org
> > > Subject: Re: MIB EDU training in Montreal
> > > 
> > > 
> > > On Wed, May 17, 2006 at 06:35:44AM +0200, Wijnen, Bert
> (Bert) wrote:
> > > 
> > > > OK, My material that I have is the slides that I presented to 
> > > > the IETF WG chairs training session at IETF60.
> > > 
> > > I think this is a very good start.
> > > 
> > > I believe it is important to add some slides which point
> out that it
> > > is very helpful to have an information model which
> explains how things
> > > are related before writing MIB modules. Perhaps it helps
> to show some
> > > examples to illustrate the point.
> > > 
> > > A simple to understand example perhaps is the Printer-MIB
> (RFC 3805)
> > > since everybody knows what a printer is. Figure 2 in RFC
> 3805 defines
> > > a conceptual block diagram for a printer and the MIB tables are 
> > > organized according to this block diagram.
> > > 
> > > /js
> > > 
> > > -- 
> > > Juergen Schoenwaelder		    International 
> University Bremen
> > > <http://www.eecs.iu-bremen.de/>	    P.O. Box 750 561, 
> > > 28725 Bremen, Germany
> > > 
> > 
>