[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Scope of NIM




>>>>> remoore  writes:

remoore> I think that most of the early discussions of "scope" were
remoore> centered on (2) - for example, Andrea's questions about the
remoore> scope of what NIM would be modeling.  This is also what
remoore> people meant when they talked about helping different WGs
remoore> model common elements in a common way.  (It's clear that this
remoore> is what they meant because they contrasted this goal with
remoore> what happened with SNMP: many WGs, all using SMI as their
remoore> data representation, came up with divergent ways of modeling
remoore> the same things.)

I certainly do understand and agree with the goal to reduce the number
of different ways to do similar things. If we try to attack this
problem, then the question is which tool we use. One tool is of course
defining a very abstract and generic model and to "force" everyone to
derive/comply to it. I personally have doubts that this works for a
variety of reasons.

Another tool which I believe can work much better (and which we failed
to implement in the SNMP world yet) is to define "design pattern" for
typical problems and to encourage people to use them. We already have
proven design pattern in the MIB world - but they are usually not
spelled out and only known by a few "insider" people.

The IESG MIB review process did help to some extend to get more
commonality into MIBs. But I believe that more can be achieved by
really writing down explicit design patterns that others can pick up
early and use to simplify their work on structuring their information
and/or data models. The BCP document being worked on in the snmpconf
WG is probably an SMI specific start into that direction.

In short:

(1) I believe that we need to look at a new data definition language
    which (a) must at least reduce the total number of languages we
    use right now from n to n-1 and which (b) prevents others from
    inventing n+1, n+2, and so on.

(2) I believe that we need to have strong usage guidelines for this
    new data definition language and a set of design pattern on how
    to use the language effectively for common problems.

(3) I do not believe that we will ever reach a commonly accepted core
    information model which crosses several technology domains (which
    is more or less equivalent to multiple IETF WGs) or even all IETF
    protocols ever defined. (I have stated my reasons for this opinion
    in other emails - so I won't repeat them here.)

I think this pretty much sets the scope as I see it.

/js

-- 
Juergen Schoenwaelder      Technical University Braunschweig
<schoenw@ibr.cs.tu-bs.de>  Dept. Operating Systems & Computer Networks
Phone: +49 531 391 3289    Bueltenweg 74/75, 38106 Braunschweig, Germany
Fax:   +49 531 391 5936    <URL:http://www.ibr.cs.tu-bs.de/~schoenw/>