Re: Verbs Again (was RE: draft-shafer-netconf-syslog-00.txt)

[Andy Bierman <ietf@andybierman.com>]

Phil Shafer wrote:
> Andy Bierman writes:
> > You are okay with standard parameters for syslog delivery
> > when they are passed in a special RPC, but if those standard
> > parameters are moved to a data model, suddenly you can't
> > map your proprietary mechanism to the standard anymore.
>
> Sure.  I think mapping to an abstract model is reasonably
> easy (think SNMP MIBs), but mapping from an abstract model
> to device configuration is not easy (think SNMP MIBs).

Think how unusable SNMP would be if the MIB objects
didn't have OID assignments and therefore every NMS
has to hard-code the vendor-assigned OID for every MIB
object, for every version of every device for every vendor.

Think how unusable SNMP would be if there was no MIB-walk,
and instead the NMS had to hard-wire the special PDU to use
for every MIB object, for every version of every device for
every vendor.

That's not exactly how standards are supposed to work.


> > So it is
> > not a burden for a conforming netconf agent or manager
> > to use the base operations, since they are never optional.
>
> The point is that <get-config> and <edit-config> for device-specific
> configuration is simple.  Doing them for a data model that is not
> implemented verbatim on the device is not.  I think you imagine
> people doing verbatim implementations of whatever data model that
> netconf defines.  I don't.  I see a big network with lots of deployed
> boxes and lots of configuration data that isn't doing to move to a
> brave new world in my lifetime.  I want to make something that's
> useful in my lifetime, so mandating this isn't doing it for me.


It doesn't matter what proprietary stuff you have on your router.
There are zero standard data models for NETCONF.  For new
NETCONF features, you can expect new standard data models
to configure them. We go from zero to one, then two, etc.
We get WG consensus on the contents of every standard data model.

This is no different than standard MIB modules.
IETF WGs have managed for years to define standard MIB objects,
even though proprietary MIB objects also exist.  With SNMP
you can use the same PDUs to access vendor data as standard data.



> > The bar has been raised in this WG.
>
> The goal is to make something that solves today's problems in
> reasonably efficient, reasonably cost-effective ways, that
> vendors will implement and operators will use.


There are many people (including myself) who believe that
standard configuration data models are a critical component
to a complete standards based solution for network configuration.

This will take time, but standard protocols should be configurable
in a standard way.  Vendor extensions to standards will always
exist, as well as proprietary solutions.

> Thanks,
>  Phil

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/>