[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: no NETCONF WG meeting planned for Paris



David B Harrington wrote:

Hi,

To me, a protocol with no data modeling language and no data models is
close to worthless as a standard. It provides only a "bucket
transport" protocol where the contents of the bucket are proprietary.



There are some pretty powerful features within a decent transaction model to act on that bucket of XML content. I know some of us on the design team we're really just trying to eliminate screen-scraping the CLI with 1.0. Even if NETCONF is just used as a programmatic interface to the CLI at first, this is better than screen-scraping.

We've had a mix of proprietary and standard SNMP data from
day one, and it will be no different for NETCONF.

I agree the netconf WG needs to focus on the protocol issues,
motivating independent implementations, and starting discussions on
partial locking and other issues that were punted to get 1.0 out the
door. I support Andy and Simon in this.

A Netconf SMI is necessary to allow the development of standards-based
data models, and a NETMOD SMI should be done in a separate WG from the
protocol. If the Netconf WG decided to not try to start a NETMOD WG, I
think that is a mistake.



I don't remember the WG agreeing to work on specific features in the future,
just that work should continue in the WG on protocol features. I do remember
that we were very concerned not to get in the same "SMING/EOS" deadlock,
where each WG wouldn't do work the other one needed.


I'm not against doing the NETMOD work in the NETCONF WG if that's
the WG consensus.  I don't see a lot of energy and momentum around the
current proposal based on the lack of email comments on the NETMOD list.

I recommend starting with an SMI that can map to existing SNMP
information, since there is a wealth of data models already available,
researchers have already developed tools that automate the
translation, and Netconf implementations could use these data models
to demonstrate that the Netconf protocol works properly. SNMP and
SMIv2 are showing their 1980's roots, so we should work on defining a
Netconf SMI in XML to provide additional capabilities, such as nested
data structures. As the SMIv3 effort showed, the SNMP-adapted ASN.1
SMI has too much baggage to be able to design a backwards-compatible
SMIv3 easily. Developing a new SMI in XML will give us a better chance
of advancing the state of the art in NM data modeling.



I think you are proving my point about the one-size-fits-all myth.
I agree that this could be one of several data modeling approaches
that can be mapped to the NETCONF protocol.


I also believe that we should start work on SNMPv4 - a version of SNMP
based on XML messaging rather than ASN.1 messaging, leveraging
existing transport layer message security, using an XML-based data
modeling language, and using an access control method that is easier
to read than VACM's OID hierarchies and can be utilized by SNMP and
Netconf, thus sharing many ease-of-use design decisions between the
monitoring and the (separate) configuration protocol standards.



This sounds interesting (do you mean s/messaging/encoding/ in line 2?)

David Harrington
dbharrington@comcast.net



Andy



-----Original Message-----
From: owner-netconf@ops.ietf.org [mailto:owner-netconf@ops.ietf.org] On Behalf Of Sharon Chisholm
Sent: Thursday, May 26, 2005 2:06 PM
To: netconf
Subject: RE: no NETCONF WG meeting planned for Paris


hi

Actually, my remembrance is that a number of items were dropped solely as a
scoping exercise. Note because they were not felt to be achievable


or


necessary.

If there is working group consensus to take on additional work and the AD is
sufficiently convinced of its value and achievability, are we then all on
board?


Just a note that the current Netmod work provides a framework to define
content but does not actually define any content.


Sharon

-----Original Message-----
From: Andy Bierman [mailto:ietf@andybierman.com] Sent: Thursday, May 26, 2005 12:19 PM
To: Chisholm, Sharon [CAR:5K50:EXCH]
Cc: netconf
Subject: Re: no NETCONF WG meeting planned for Paris



Sharon Chisholm wrote:



hi

So, as I indicated in Minneapolis, I plan two presentations into


the

Paris meeting. One on the Netmod work proposing that gets

added to the

charter and another on asynchronous messaging with a similar objectives. The technical work is being progressed as we speak.

Are you suggesting that you would like to see the proposed updated charter before Paris?




Simon and I both agree that the NETCONF WG should concentrate on finishing
the 1.0 document set and then take a little break -- like until the RFCs are
actually published -- before charging ahead with more features. We would
like to make sure the IESG approves the work we've done so far


before


building on that work. A little time for implementation, inter-operability
testing, and deployment would be a good idea as well.


Before this new work, we would of course need to update the WG charter, and
to do that, we need consensus on that charter. Neither Simon or myself is
eager to take on the NETMOD work in its current form. The deferred protocol
features like notifications can be started if there if WG consensus.


Of course, every feature on that list is there because it is difficult, contentious,
and the WG couldn't reach agreement on it. I'm not convinced we will do any
better the second time around. (Prove me wrong -- write a great draft that
wins the WG over.) Solution proposals full of the same holes as the first
time around aren't going to become WG drafts just because they


exist.


More on NETMOD: My preference is for NETCONF to be content-neutral,
like the document claims. I would like to see other SDOs and even


a

NETMOD
WG in the IETF define mappings between their data modeling constructs and
the NETCONF protocol (operations and transaction model). I do not think
a one-size-fits-all solution will ever work, or that the first cut at a data modeling
standard will be long-lasting. Personally, I haven't seen anything yet that really
removes enough complexity to be interesting, but I'm confident in


the

next few
years somebody will figure it all out.



Sharon




Andy



-----Original Message-----
From: owner-netconf@ops.ietf.org


[mailto:owner-netconf@ops.ietf.org] On

Behalf Of Andy Bierman
Sent: Tuesday, May 10, 2005 12:08 PM
To: netconf
Subject: no NETCONF WG meeting planned for Paris


Hi,

At this time, it appears all WG milestones will be completed

before the

next IETF, so there are no plans to hold a NETCONF WG meeting at


the

Paris IETF.

We understand that some people want to continue adding

features to the

protocol and/or work on standard data models, but this work

would need

to be chartered first -- a process that has not been

started, and not

likely to be completed before the next IETF.


Andy and Simon

--
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/>


--
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/>







-- 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/>






-- 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/>






--
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/>