RE: What about the OAM in OAM&P

["Hare, Michael" <Michael.Hare@us.fujitsu.com>]

Thanks Andy.

The Notifications is what I think was confusing some of us into believing NETCONF was to be a complete OAMP solution since notification have little to do with network configuration.

From what I gather from your response NETCONF is just for the configuration of the network with a bit of notification thrown in.

Maybe there's a need for other WGs to define NETMAN, NETOP and NETADM to cover the rest. Not your problem, I understand, but I have to convience the folks here that things like software upgrades, loopbacks and such are not part of NETCONF. We will need to write proprietary RPCs for all that.

Thanks,

- m

-----Original Message-----
From: Andy Bierman [mailto:ietf@andybierman.com]
Sent: Monday, July 03, 2006 11:29 AM
To: Hare, Michael
Cc: Netconf Mailing List (E-mail)
Subject: Re: What about the OAM in OAM&P


Hare, Michael wrote:
> Hi,
> 
> Please forgive me if this was covered.
> 
> Netconf is doing a good job of covereing the 'P' and maybe some of the 
> 'A' and 'M', but what about the 'O'?
> 
> Things like:
> 
> Software Download (not really a <copy-config>, maybe just <copy>?)
> Operating and Releasing Loopbacks/Test Signals (not really editing the 
> configuration, new verb-set?)
> Initializing PM Registers (again, not really editing the configuration)
> Protection Switching
> 
> 
> This may be outside the Netconf charter, but it seems the rpc framework 
> being used for netconf could be re-used to the other parts of network 
> management.

My position on NETCONF extensions is clear:

Vendors should develop extensions on their own, in running code,
and get some operators to use them.  When you have some
deployment experience with the extension, then propose it
as a standard.  In my experience, if a "feature" is important
enough, vendors will provide it to customers, regardless of any standard.

<config-rant>

Notifications got grandfathered in because they were in the original charter.
After that, we will be working on standards based configuration
or nothing at all.

Our focus is standards based configuration, a subject some
vendors seem to want to leave as "TBD".  Anybody who understands
embedded agent design knows why --  it is a non-trivial matter
to support multiple incompatible configuration APIs
(e.g., proprietary and standard), unlike multiple incompatible
monitoring APIs.  The problem is not "special RPC" vs. "data model" at all.


Unfortunately, the current process and mindset for IETF data models (MIBs)
is never going to work for configuration.  Somebody is
going to have to figure out how to remove 1 of the 2 adjectives
('multiple' or 'incompatible').  Maybe this is a topic for the NMRG.

</config-rant>


> 
> 
> ---------------------------------
> Michael Hare
> NMI Engineering

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/>