design: special RPC vs. data model

[Andy Bierman <ietf@andybierman.com>]

Hi,

I think the WG should understand and write down the
architecture for netconf wrt/ its use within a
network management system (not just netconf, not just 1 device).
That is not our charter at this time, so we don't
have to write an entire document, but we do have to consider
the overall architecture in the notification work.

For starters, we already have a 'traditional' model
where a manager manipulates configuration databases
on the agent, using standard RPC methods like edit-config.

Although specialized RPC methods like create-subscription
are always possible in netconf, they are not always the
best design choice.  We don't seem to have any criteria
for deciding whether to use custom RPC methods or a data model
manipulated with the standard RPC methods.

The current draft has custom RPC methods, plus a mandatory
read-only data model for use with the standard method <get>.
I don't understand this architecture.  If the manager is
sending session-specific parameters (and that's why a
custom RPC is appropriate), then why do we need a global
data model for others to read back the settings
in use for all the various sessions?   If these are
not session-specific parameters, then why not use the
traditional read-create data model + standard methods approach?


Andy


--
to unsubscribe send a message to netconf-request@ops.ietf.org with
the word 'unsubscribe' in a single line as the message text body.
archive: <http://ops.ietf.org/lists/netconf/>