[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: Charter items
At 05:51 PM 3/28/2003 -0500, RJ Atkinson wrote:
>On Friday, Mar 28, 2003, at 17:21 America/Montreal, Ron Bonica wrote:
>>If we were to standardize a few small chunks of configuration and state
>>information, we could convince ourselves that the worst problems can be
>>solved within the context of XMLCONF. I'm not saying that this working group
>>should translate every MIB ever developed by the IETF. Maybe we should limit
>>the effort to codifying the configuration and state data that currently
>>resides in the MIB-II ifTable. It should also document best practices for
>>defining XML configuration and state data.
>
>Ron,
>
> In my mind, it isn't a question of what to do, so much as a question
>of what to work on first. I'd prefer to work on how to move opaque config
>blobs around first, which gets rid of screen scraping, then move on to
>other stuff (in my mind, "other stuff" is pretty open-ended at this point).
>
> From the above, I gather you prefer to work on everything at once.
>My own past experience is that IETF WGs are not at all successful when
>a WG tries to work on more than one thing at a time. It sounds like
>your mileage might vary from mine.
>
> Reasonable people might well reach different conclusions on
>this topic.
I agree with you about limiting the initial phase of work.
SMI and data model work can either wait or be done by another WG.
(SMIv2 to XSD conversion work has already been done by the libSMI team.)
However, there are a few areas of overlap which cannot be ignored
by this WG:
1) XML usage guidelines
The XMLCONF draft recommends the usage of Canonical XML (RFC 3076).
This is somewhat controversial. In general, equipment vendors
want to support a subset of full XML in network devices, since
some XML features (e.g. include directives) are expensive to
support in embedded systems.
2) protocol syntax specification
The XMLCONF draft uses XML Schema Documents to define the syntax
of protocol messages. It is assumed that data model definitions
will also use XSDs. Whatever is selected for the SMI must be
consistent with the protocol document.
3) namespace definitions
The XMLCONF draft makes some assumptions about the structure of
namespace URIs (e.g., http://ietf.org/xmlconf/1.0/base). The
SMI and standard data model namespace definitions need to be
consistent with the structure defined in the protocol document.
I think the netconf WG can proceed with a protocol definition,
even with these coupling issues.
>Cheers,
>
>Ran
>rja@extremenetworks.com
Andy
--
to unsubscribe send a message to xmlconf-request@ops.ietf.org with
the word 'unsubscribe' in a single line as the message text body.
archive: <http://ops.ietf.org/lists/xmlconf/>