RE: netconf WG charter proposal

["Faye Ly" <faye@pedestalnetworks.com>]

Andy,

Certainly and I agree that is a good approach to get the most needed
stuff done first.  I do argue against the monitoring/notification
channel being done with this group as I don't think it can be achieved
without data modeling.  Take the existing 'trend analysis' tools for
example, it has to do inventory first to discover the interface that is
worth trending.  Different vendors can name their interfaces quite
differently and this could be a nightmare to try to figure out if one
interface from one box is actually symmetric to another interface from
another box. (They are physically or logically connected and thus need
to be trend and compared accordingly).

-faye


-----Original Message-----
From: Andy Bierman [mailto:abierman@cisco.com] 
Sent: Monday, April 14, 2003 4:08 PM
To: Faye Ly
Cc: David T. Perkins; Chen, Weijing; xmlconf@ops.ietf.org
Subject: RE: netconf WG charter proposal

At 02:59 PM 4/14/2003 -0700, Faye Ly wrote:
>Andy,
>
>If XML monitor/notification is standardized with configuration, yes.
Or
>else if we are trying to identify which SNMP notification matches to
>what interface or instance in the configuration, no.  Ethernet0/0 is
>still a simple case.  Take a look at the CIDR table and also table
>indexed by DisplayString.
>
>I honestly think that having to focus on the configuration only is a
>well defined goal for this discussion group and standardization can be
>achieved.  If this group is trying to do all three, XML SMI seems to be
>un-avoidable.  What do you think?

This WG is not trying to create standard data models or a new
data modelling language.  That will need to be done at some point,
probably by a different WG.  The protocol does not rely on any
particular data model design.  We will attempt to identify any 
coupling between the protocol and data modelling capabilities,
but not design a new SMI.


>-faye

Andy



>-----Original Message-----
>From: Andy Bierman [mailto:abierman@cisco.com] 
>Sent: Monday, April 14, 2003 1:21 PM
>To: David T. Perkins
>Cc: Chen, Weijing; xmlconf@ops.ietf.org
>Subject: RE: netconf WG charter proposal
>
>At 01:02 PM 4/14/2003 -0700, David T. Perkins wrote:
>>HI,
>>
>>Syslog... Let's see, the payload of syslog messages is unstructured
>>text, and there is no standard for them. And there is no standard
>>for documenting the payload. Or did I miss this. (If so, please
>>send me a URL or two). So you are saying that completely unstructured
>>and undocumented text is the direction for the future?
>
>I don't think Weijing is suggesting that we use unstructured text 
>instead of XML encoding.  Look at the example in his previous mail:
>
>    <rpc-notif>
>        <cause>linkDown</cause>
>        <source>
>                <interface>Ethernet0/0</interface>
>        <source>
>    <rpc-notif>
>
>He is saying (and I agree) that the XML data model does not
>have to use ifIndex to identify the interface.  The agent
>probably needs to know (internally) the ifIndex mapping for 
>Ethernet0/0, but that doesn't mean that the XML data model 
>must use it.
>
>People have been incorrectly assuming that the XML data model 
>will need to be tightly coupled to the SMI data model.  
>
>Andy
>
>
>
>
>>At 02:38 PM 4/14/2003 -0500, Chen, Weijing wrote:
>>>Why do the mapping at all?  Why can we ditch the SNMP altogether,
>>>configuration through XML, trap through syslog over XML, or
structured
>>>native XML traps over XML.  There is no much detailed trap
information
>from
>>>SNMP at the beginning.  All we got from most of equipments are
>linkUp/Down,
>>>etc.  Most of equipments have more detailed information from syslog
>rather
>>>than SNMP trap.
>>
>>Regards,
>>/david t. perkins 
>>
>>
>>--
>>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/> 
>
>
>--
>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/>
>
>--
>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/> 


--
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/>