Re: Incomplete XML Draft

[Andy Bierman <abierman@cisco.com>]

At 12:39 PM 6/25/2002, RJ Atkinson wrote:

>On Tuesday, June 25, 2002, at 03:18 , Andy Bierman wrote:
>>Some people think syntax without semantics is too narrow a charter,
>>and not that useful.
>
>I wasn't aware that even syntax was on the table.  The notion that
>some of us had put forth as an initial goal was simpler than that:
>        - standardise a way to move proprietary XML to/from switches and routers
>
>>We already have standard protocols for
>>moving arbitrary data around.
>
>...which many operators don't believe are useful for configuring network equipment.
>
>>The charter that Ted proposed makes much more sense, and will result in some amount
>>of standards-based configuration management.
>
>...which will likely be as widely ignored by operators as the current SNMP approach,
>because it doesn't solve the problems that many operators are saying they have today.
>
>I have this wacky idea that the IETF ought to tackle the OPS/NM problems that
>many operators say they have -- rather than tackling the OPS/NM problems that
>IETF folks think are interesting but lack widespread current operator interest.
>Please forgive me for being consumer oriented.

I'll wait for the BOF to see what all the attendees think.
I don't think your opinions represent those of all network operators.


>>I'm not proposing this as a replacement for SNMP.  I am proposing
>>that the charter include mapping mechanisms between data defined
>>in MIBs and XML encoded protocol operations.
>
>I don't see how that helps with configuring anything but a DOCSIS CM
>-- and I doubt seriously that the DOCSIS operators will be changing
>their provisioning systems anytime soon.

There are plenty of proprietary MIB objects for configuration.
There are even some standard MIB objects for configuration.


>>I'm not interested in reinventing every knob that's ever been defined
>>in a standard or proprietary MIB.  I also think it will be useful
>>for applications to recognize and translate names so monitoring
>>data can be processed -- e.g. what interface is associated with
>>ifInOctets.42?
>
>I'm not even thinking about monitoring yet; box configuration
>is the *only* target in *my* mind for the near-term.   Clearly
>your mileage varies.

I'm thinking about an application which wishes to monitor with SNMP
and configure with XML. It needs to know how to translate the '42'
in ifInOctets.42 to an interface name in XML.


>>>I've seen too many WGs with overly broad charters fail to wish that
>>>death on any more BOFs/WGs.
>>
>>I don't think the addition of "SMI --> XML name translation" will kill
>>the WG. (The XML content should not be limited to translated SMI,
>>satisfying those who just want a standard way to move proprietary
>>config data around.)
>
>I wouldn't object to it as "possible future work" -- provided the IESG
>keep it VERY explicitly out of scope for an XMLconf WG until some
>MUCH MUCH simpler stuff (e.g. moving XML to/from network boxes) has been
>visibly successful.
>
>Perhaps what you and others might consider is going to the IESG and proposing
>a *separate* WG on "SMI-->XML name translation".  I don't know what the IESG
>would say, but asking can't hurt (much).

I'll abide by the majority consensus, as determined at the BOF.
If the majority thinks SMI->XML name translation should be part
of a separate WG, then that's fine with me.


>Regards,
>
>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/>