[NGO] comments on CANMOD BoF

Hi,

I was encouraged to hear that some consensus on a path forward
with YANG as the high-level DML, and standard XML languages
as the low-level DMLs.

I am still concerned about the absence of 'coexistence with
SNMP MIBs' in the game plan.  I have a common and important use case:

What if a WG wants to create a NETCONF MIB module to configure
a protocol, which has existing detailed capabilities already
standardized in an SMIv2 module?  What if the WG wants machine-readable,
cohesive mechanisms to indicate that certain config features/knobs
are only available or applicable if certain MIB objects have particular
values?

Randy's answer at the Bof was that WGs will just have to
use SNMP, not NETCONF, to configure their protocols.
I strongly object to this solution approach, and expect the
NETMOD architecture to fully account for SMIv2 objects, TCs,
and notifications, plus SYSLOG notifications, and standard
RPC operations added by other WGs.

I also expect the WG charter to include a coherent and robust plan
for operational security, that includes a standard access control model.
This must fully account for all aspects of NETCONF protocol behavior,
including vendor actions which interact with the same configuration
data as standard operations.

Anything less would be a toy, not suitable for real networks.

Andy

_______________________________________________
NGO mailing list
NGO@ietf.org
https://www.ietf.org/mailman/listinfo/ngo