[netmod] Approval of minutes 20100820: Draft minutes (v1) of netmod working group session IETF 78

[David Kessens <[e-mail]>]

Please see attached for our first version of the minutes.

We would like to declare the minutes final if we don't receive
any comments (other than editorial corrections) by August 20, 2010.

Thanks,

David & David
---

Draft minutes (v1) of netmod working group session IETF 78

Meeting:   IETF 78
Location:  MECC, Maastricht, Netherlands
           TUESDAY, July 27, 2010, 09.00-11.30, Meeting Room: 0.9 Athens
WG Chairs: David Kessens (david.kessens at nsn.com)
	   David Partain (david.partain at ericsson.com), not present
Jabber:    xmpp:netmod at jabber.ietf.org
WG URL:    http://tools.ietf.org/wg/netmod/
Minute taker: Andy Bierman

The working group session started with an status update compared to the
netmod working group charter and on all relevant drafts.

The working group is a little late compared to the charter but most of the
work is now at the IESG level or beyond so the working group is getting
close to completion. The chairs expressed their happiness with the work
done so far and thanked all participants for getting the work so close to
the finish line.

The status of the following working group drafts were discussed:

1) Common YANG Data Types
   http://tools.ietf.org/html/draft-ietf-netmod-yang-types-09
   Status: RFC Ed Queue
   (Juergen Schoenwaelder)

   no further updates from the authors or comments
       
2) YANG - A data modeling language for NETCONF
   http://tools.ietf.org/html/draft-ietf-netmod-yang-13
   Status: RFC Ed Queue
   (Martin Bjorklund)
 
   no further updates from the authors or comments
 
3) YANG Usage Guidelines
   http://tools.ietf.org/html/draft-ietf-netmod-yang-usage-09
   Status: Waiting for AD Go-Ahead
   Slides: http://www.ietf.org/proceedings/78/slides/netmod-4.pdf
   (Andy Bierman)
   
   Andy Bierman discussed the current status and updates since last meeting:
   
   - updated details from Gen-ART review
   - should security guidelines be online or in doc
     [ put current text in draft; pointer to WEB;
       note that WEB is normative]
   - mandatory=false; config=false
     [ put note that data model writer needs to clarify
       conditions for server implementation; object
       semantics may make it clear what to do; if not
       text should be present to clarify conditions;

   3.4 NETMOD Architecture
       http://tools.ietf.org/html/draft-ietf-netmod-arch-07
       Status: IETF Last Call Requested
       (Phil Shafer)
       
       no further updates from the authors or comments
   
   3.5 Mapping of YANG to DSDL
       http://tools.ietf.org/html/draft-ietf-netmod-dsdl-map-06
       Status: AD Evaluation
       Slides: http://www.ietf.org/proceedings/78/slides/netmod-0.pdf
       (Ladislav Lhotka)
       
   Ladislav Lhotka discussed the current status and updates since last
   meeting:

   - WGLC and external review changes
     - list of terms added
     - terminology cleanup
     - simplified the mapping procedure
     - implementation in pyang tool (yang2dsdl) is now up
       to date with the draft 
    - DSDL tutorial added
     - see www.yang-central.org for details
     
   - Issue brought up by Dan Romascanu: there is a normative reference to an
     Informational RFC - needs to be changed/dealt with by calling it out in
     Last Call message

After the working group drafts, the status of other individual drafts
relevant for the netmod working group were discussed:

1) draft-bierman-netconf-system-monitoring-00.txt
   Status: new submission
   (Andy Bierman)

   Draft was discussed in the NETCONF working group.
   The working group chairswanted to make sure that everybody is aware of
   this work.

2) draft-linowski-netmod-yang-abstract-03.txt
   Status: headed for AD sponsored experimental
   (Bernd Linowski)  

   This draft was discussed during the Ops Area working group session.
   This work was not part of the current netmod charter and was accepted
   by the Operations and Management Area Director to be considered and
   published as an experimental RFC.
   
   The chairs asked for volunteers for a review of this document to help
   Dan. Martin Bjorklund volunteered to be Dan's reviewer.

3) draft-schoenw-netmod-smi-yang-00.txt
   Status: expired, but possible work item for new charter
   Earlier presented slides: 
   http://www.ietf.org/proceedings/75/slides/netmod-0.pdf
   Slides this meeting: http://www.ietf.org/proceedings/78/slides/netmod-1.pdf
   (Juergen Schoenwaelder)

   Juergen Schoenwaelder discussed his draft in the context of accepting this
   work as a future work item for a rechartered netmod working group:

   - mapping SMIv2 modules into YANG
     - code is part of libsmi
     - allows access to SNMP instrumentation via NETCONF
     - direct translation
     - code rewrite will allow more conversions
     - do we need a translation for SMIv2 to YANG

A large part of the meeting time was allocated to discuss the next steps for
the netmod working group. Whether is is decided to recharter or close a new
working group, the chairs proposed to discuss what work still remains to be
done now that netmods scheduled work items are mostly complete.

David Kessens presented the following presentation:

http://www.ietf.org/proceedings/78/slides/netmod-3.pdf

The slides presented a number of questions that needed discussion or answers
before a new proposed charter can be contructed.

The first discussion item that was brought up was whether the working group
should close down or continue with a fresh new charter.

   - Area Director: should use this meeting time to
     discuss recharter now
     need to continue momentum and add YANG modules
   - David Kessens: last meeting preference seemed to be for recharter

The following topic was on what issues still need to be solved in the IETF
to make YANG deployment happen. The working group was invited to comment.

   - Lada: improvement to data type system
       - native IP address instead of patterns
         - how better: string compare can cause
           same addess in different forms not to match
         - string compare should work because the server
           will normalize 
       - objection to using extensions to build the
         standard;
   - Martin: what about network-wide-config; is that
      in scope? 
   - Andy: need to deal with SNMP data structures being referenced
      from NETCONF (they do not exist in NETCONF; assumption
      that SNMP agent will be present)
     - system core module is needed
     - interface list should be core module
     - refinements to the conformance model over time
     - mapping from SMIv2 to YANG is not exactly the
      same as mapping SNMP to NETCONF (e.g., contextId
      is part of SNMP, not SMI)

Slide 3 was about what work should be explicitly excluded and the floor was
opened for comments:
      
   - no desire to start YANG 2.0 now
   - network-wide config may require changes to YANG
     - framework for net-wide-config may lead to lots of
       of different work; not clear yet what this will be
   - need for YANG Doctors review team, but not done in
     the NETMOD WG
   - operational state?  not in the charter but will come up
   - config alternative states (pre-provisioning, disabled 
     config, etc.)

Slide 4 asked for specific work items and whether there were volunteers to
work on them:
   
    - SMI to YANG mapping (Juergen)
    - network-wide config (Phil)
    - routing table data model
      - SMIv2 is horrible for representing a routing table
      - need YANG model that can be extended with protocol-specific
      properties

Finally, Andy and Juergen volunteered to work on a new charter text together
with David Kessens based on the comments received during the previous
discussions.

The question on whether a new working group should be started or a
recharter was better was brought up again.

    - AD: slightly easier to recharter than start a new WG
    - David Kessens (wgchair): prefer to close and charter a new WG, 
      but no objection to writing charter text assuming a recharter

Also, the topic of net-wide config was brought up again:
  - need to see more details to know what the charter text 
    - not really needed
    - David asked the working group: should be included?
       - 1 OK; 3 no; 3 leave in the middle; 14 don't care

Under the "Any other Business" agenda topic, Dan Romascanu brought up the
issue of a YANG module review team. He made it clear that he believes a YANG
doctor group should be setup and he announced that he was going to create a
mailing list and announce a call for volunteers for the team.

One issue was brought up for the architecture draft, whether the data
distinction problem described (operational vs. config) needs 2 leafs for
operational vs. config.
   - Martin agrees to add; Andy agrees; no objections.
   
=============================================