RE: netconf WG charter proposal

[Randy Bush <randy@psg.com>]

</ad-hat>

apologies for poorly articulated uneasiness.

> The XMLCONF draft does contain separate <get-config> and
> <get-state> protocol operations.  This was done to satisfy an
> operator requirement (identified at the NM Workshop) that it
> should be easy to retrieve config data in a format that can then
> be directly fed back to the device in a 'set config' operation.

because one asks for A does not directly imply that one also wants
B.

> IMO we need the <get-state> operation to retrieve non-config
> data, such as interface counters, the device uptime, the list of
> users currently logged in, etc.

i agree that i want those data.  and i want to reset counters too.
i am slightly unsure that i want to insist on the same protocol and
language to do this as i do for configuration.  it's not that the
elegance does not appeal to me.  it is that i am not sure i will
like all the consequences.

[ please be very aware that i am very intentional in my use of
  words such as 'unsure', 'suspect', etc.  i am really groping here
  trying to understand why i am uneasy. ]

i am not here to bury snmp, but to praise automated configuration.

i am worried about distractions, complications, ratholes, and goal
drift if i start to walk the 'state data' path.  and my fears could
well be unjustified.  

i suspect that marshalling state data in xml will not seriously
complicate the schema from a pure config schema.  so maybe it is
uneasiness that the transport, transaction, authorization, ...
models will be more complicated?

randy


--
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/>