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

RE: Methods in the NIM requirements

I know that Juergen said that he didn't want to discuss this - but why bring
up an example and then say "don't discuss this"?  The DMTF teams (both
Device and Networks) really did give some thought to the Speed property in
CIM_NetworkAdapter.  Here is the background:

1.  "Speed" was modeled as a uint64 since it was discussed that maximum
growth range should be allowed and that units should stay "bps" for
consistency with the IETF MIB.
2.  If the Speed value actually fits in a uint32, then the mapping is clear.
If the Speed is greater than a uint32 with units of bps, then reality can't
be reflected in RFC1213's attribute.
3.  The goal was to map both the MIB and DMI, where DMI defined the property
with units of Mbps.


> -----Original Message-----
> From: owner-nim@ops.ietf.org [mailto:owner-nim@ops.ietf.org]On Behalf Of
> Juergen Schoenwaelder
> Sent: Tuesday, May 02, 2000 2:18 AM
> To: jstrassn@cisco.com
> Cc: nim@ops.ietf.org
> Subject: Re: Methods in the NIM requirements
> >>>>> John Strassner writes:
> John> No one is ignoring the MIBs and other protocols. If you would
> John> actually look at one of the models, you would realize that it
> John> already maps some MIB information. This was true of CIM and DEN
> John> 2 years ago, and is true of the Policy models that have been
> John> recently published as IETF drafts.
> The CIM MIB mappings I have seen are often broken and fail to achieve
> the goal of unification and overall simplification. An example from
> the CIM Devices MOF specification 2.1.1:
>         [Description(
>           "An estimate of the current bandwidth in Bits per Second. "
>           "For endpoints which vary in bandwidth or for those where "
>           "no accurate estimation can be made, this property should "
>           "contain the nominal bandwidth."),
>          Units ("Bits per Second"),
>          MappingStrings {"MIB.IETF|RFC1213-MIB.ifSpeed",
>                "MIF.DMTF|Network Adapter 802 Port|001.5"}
>         ]
>     uint64 Speed;
> Note that the original definition of ifSpeed is a 32 bit gauge and not
> an uint64. (This document also refers to BRIDGE-MIB.dot1dTpPortMaxInfo,
> which is defined as an INTEGER in the MI, but CIM defines it to be an
> uint32.) [This may have all been done on purpose - but there is no
> text which explains how you handle the interesting corner cases that
> may show up in practice.]
> I do not want to discuss this example (or others). I just want to warn
> everyone that mappings are (a) not simple and (b) require that domain
> experts are involved and (c) take time to get right (which requires to
> have mechanisms in place to evolve NIM models over time without
> causing interoperability problems).
> /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/>