Minutes IETF106: netmod
minutes-106-netmod-01

Netmod IETF 106 Minutes

Session 1        "10:00-12:00
YouTube:        https://www.youtube.com/watch?v=nqBsv6pNuZk

Presentation                Start Time        Duration        Information
1                10:00        10        Title:        Administrivia & WG Status
                                Draft:
                                Presenter:        Chairs (Lou Berger, Local,
                                Kent Watsen, Remote)

2                10:10        5        Title:        WG LC Update: Interface
drafts
                                Draft:       
                                https://tools.ietf.org/html/draft-ietf-netmod-intf-ext-yang-08
                                        https://tools.ietf.org/html/draft-ietf-netmod-sub-intf-vlan-model-06
                                Presenter:        Rob Wilton

Acee Lindem: Clients could determine the in_pkts and out_pkts counters by
aggregating the unicast, multicast, and broadcast counters together

Rob Wilton: Agree, but this doesn't work if you cannot classify the traffic
into those buckets, and it can make xpath expressions that want to look at the
total values more complex.

3                10:15        10        Title:        YANG instance-data-format
                                Draft:
   https://tools.ietf.org/html/draft-ietf-netmod-yang-instance-file-format-04
                                Presenter:        Balazs Lengyel

Rob Wilton: With the YANG semver work, we may not want to use the date version.

Balazs Lengyel: Would prefer to defer to a -bis version of the draft.

Lou Berger: Personal preference would be that you cover what is supported by
current definitions and future options update this document if/when they are
defined.

Andy Bierman: Please use YANG and not YIN. 1 or 2 is ok

Kent Watsen: Prefers 2

Martin Bjorklund: Also prefers 2

4                10:25        15        Title:        DT Update overview &
status update
                                Draft:
        https://www.ietf.org/html/draft-ietf-netmod-yang-versioning-reqs-01
        https://www.ietf.org/html/draft-verdt-netmod-yang-solutions-02
                                Presenter:        Rob Wilton

Rob W: should be at the state where the whole set can be considered for working
group adoption.

5                10:40        15        Title:        Updated YANG Module
Revision Handling
                                Draft:
       https://www.ietf.org/html/draft-verdt-netmod-yang-module-versioning-01
                                Presenter:        Bo Wu

Joe Clarke: Should not aim for an exhaustive list (probably not possible).

Lou Berger: Which of the open issues need to be addressed before adoption.

Robert Wilton: None. They can be addressed during normal WG processing.

6                10:55        10        Title:        YANG Semantic Versioning
                                Draft:
                https://www.ietf.org/html/draft-verdt-netmod-yang-semver-01
                                Presenter:        Joe Clarke

Charles Eckel: What happens to models that use something that looks like
semver, but do not have/support semver?

Joe Clarke: Use of semver would be noted as They would have to mark in their
revision (revision-label)

Balazs Lengyel: perhaps use  vendor string, e.g., ericsson, to break the semver
regex to make it clear that semver isn't being used

7                11:05        20        Title:        YANG Packages
               (10:53)        Draft:
              https://www.ietf.org/html/draft-rwilton-netmod-yang-packages-02
                                Presenter:        Rob Wilton

Martin Bjorklund: Is this because we do not trust the revision-date or label.

Robert Wilton: We specify a URL from where you can get the package. The
checksum is to make sure you are getting it from the URL that is specified.

Kent Watsen: Extraction from XML (non-text) is lossless.

Martin Bjorklund: The new RFC format (non-paged) is also lossless for text.

Reshad Rehman: But the RFCs are not the only source of YANG that we care about.

Lou Berger: Important to ensure that the checksum computation approach is
defined and unambigious

Rob Wilton: Yes, and this can be done after adoption.

Rob Wilton: Do we need to define support for Tag operations now, or later.

Lou Berger: We can wait until later and see if important enough to anyone that
they contribute text.  If not, then we can proceed without it and add it
if/when it becomes important to the working group.

Bo Wu: The package concept is useful.  Text to describe how IETF should define
packages might be useful

Lada: Adding a signature to package definitions may be helpful

8                11:25        20        Title:        YANG Schema Version
Selection
                   11:31             Draft:
        https://www.ietf.org/html/draft-wilton-netmod-yang-ver-selection-01
                                Presenter:        Joe Clarke

Martin: Why a different mechanism for NC?  why not use the same mechansim as
for RC?

Joe Clarke: restconf had the ability to a different url, we thought that an rpc
seemed more natural for netconf to do.

                11:45        15        Title:        YANG Schema Comparison
                                 Draft:
       https://www.ietf.org/html/draft-verdt-netmod-yang-schema-comparison-00
                                Presenter:        Rob Wilton/Balazs Lengyel

Adjourn                12:00

Session 2:        "15:20-16:50        Tuesday Afternoon session II"
Youtube: https://www.youtube.com/watch?v=5ar2jbqlnYI

Presentation                Start Time        Duration        Information
1                15:20        5        Title:        Administrivia
                                Draft:
                                Presenter:        Chairs

9                15:25        10        Title:        YANG Geo Location
                                Draft:
            https://tools.ietf.org/html/draft-ietf-netmod-geo-location-02
                                Presenter:        Christian Hopps

Robert Wilton: IoT devices generate datatype string

Martin Bjorklund: All nodes are optional

Tim Carey: We typically allow for seconds from or a datetime. It works for
IoT and other devices

Martin: Make it a choice (statement)

11                15:38        10        Title:        A YANG Data model for
Policy based Event Management
                                Draft:
                https://tools.ietf.org/html/draft-wwx-netmod-event-yang-04
                                Presenter:        Michael Wang

Lou Berger: Do the authors of the previous documents agree with the progress on
the document?

Xufeng: It is progressing in the right direction.

Igor: Good progress. Agree with Xufeng

Benoit: focus on router use cases, don't make it too complex

12                15:59        10        Title:        Framework for Use of ECA
(Event Condition Action) in Network Self-Management
                                Draft:
                https://tools.ietf.org/html/draft-bwd-netmod-eca-framework-00
                                Presenter:        Daniel King

Lou: Reading the document it looked like a simple discussion on using ECA,
listening to the presentation it seems the document is a lot more complex than
initially thought.

Dan: Well, we think there is some complexity and it should be documented. We
can continue to work on the document for now then. This document does not
impact the ECA data model effort.

Lou: Yes, but it keep it as simple as possible. 18-pages is good, >100 pages is
bad.

Joe: It sounds like this document might make a good implementation guide.

Dan: We definitely have sections that are useful for implementation and best
practices, it may be worthwhile thinking about how the document, or content,
could be useful as an implementation guide.

10  16:10  10  Title:  Comparison of NMDA datastores
         Draft:  https://tools.ietf.org/html/draft-ietf-netmod-nmda-diff-03
         Presenter:  Alex Clemm

Charles Eckel: With the packages draft, do you see any implications on this
draft. Update to security consideration section??

Alex: A review of the Security Consideration section maybe needed??

Lou: Will confirm with my co-chairs, expect will move  to LC.

13                16:20        10        Title:        YANG Data Node Self
Explanation Tags
                                Draft:
              https://tools.ietf.org/html/draft-tao-netmod-yang-node-tags-00
                                Presenter:        Tao, Ran

Balazs: Conflicts with somethng that is defined in netconf-notifications draft.
Not good to have to two methods for doing the same thing.

Qin Wu:

Chris: The concept of module tags has been used in YANG catalog. Would help to
wait and see how it gets used. Let it bake for sometime.

Benoit: Tag at model level no. Tag at node level, yes.

Lou: Number too small for adoption, but small numbers of people have shown
interest, who should work together to bring the discussion to the mailing list.

-- Discussion on DT documents as WG Docs (10-20 min)

Lou: How many feel that the documents are moving in the correct direction? A
few. Does anyone have concerns. A smaller number.

Chris: Concern with the complexity of the scheme.

Martin Björklund: Concern that the schema-selection seems really complex

Andy Bierman:  +1 to concerns about complexity

Martin Björklund Note: complexity in version-selection. Not in all other drafts.

Lou: Support for adoption of the entire set of drafts with a caveat that a
simpler version selection process is proposed. Quite a few hands.

Chris: Have concern of adopting a draft that is a week old.

Lou: Adoption call will be made in January after Robert has had a chance to
update the draft.

-- (time permitting) generic model for server capabilities

Robert: Need more granulity.

14                16:15        10        Title:        using org-rfc for yang
model drafts (if time permits)
                                Draft:                                
                                Presenter:        Christian Hopps

> Adjourn                16:25