[pim] Robert Wilton's Discuss on draft-ietf-pim-igmp-mld-proxy-yang-08: (with DISCUSS and COMMENT)

[Robert Wilton via Datatracker <[e-mail]>]

Robert Wilton has entered the following ballot position for
draft-ietf-pim-igmp-mld-proxy-yang-08: Discuss

When responding, please keep the subject line intact and reply to all
email addresses included in the To and CC lines. (Feel free to cut this
introductory paragraph, however.)


Please refer to https://www.ietf.org/about/groups/iesg/statements/handling-ballot-positions/ 
for more information about how to handle DISCUSS and COMMENT positions.


The document, along with other ballot positions, can be found here:
https://datatracker.ietf.org/doc/draft-ietf-pim-igmp-mld-proxy-yang/



----------------------------------------------------------------------
DISCUSS:
----------------------------------------------------------------------

Hi,

My discuss covers two of comments below that I would like to have some
discussion on to resolve (further details are in my comments below):

(1) The addition/default of the "enable" leaf.
(2) Name of the interface_name list key rather than just name.

Regards,
Rob


----------------------------------------------------------------------
COMMENT:
----------------------------------------------------------------------

Hi,

Thanks for this document.

I note the formatting of the .txt file is strange (e.g., my commenting script
isn't able to parse it because it seems to have extra line breaks in unexpected
places).  I also note that there is no XML file uploaded.  So, sorry but my
comments are in a slightly more raw format!

1. Introduction:

 The "Network Management
Datastore Architecture" (NMDA) adds the ability to inspect the current
operational values for configuration, allowing clients to use identical
paths for retrieving the configured values and the operational values.

## I don't think that this paragraph is required.  Other IETF YANG RFCs just
cite NMDA.

2. Design of Data Model

 This document provides freedom
for vendors to adapt the data model to their product implementations.
For example, some vendors could support configuring IGMP Robustness
Variable under the interface which enabled IGMP Proxy. They could make
their own augmentation.

## I'm not sure that this paragraph is necessary, since this applies to all
YANG models and isn't specific to what is defined here.

2.2. Optional Capabilities

There is also value in widely supported features being standardized, to
provide a standardized way to access these features, to save work for
individual vendors, and so that mapping between different vendors'
configuration is not needlessly complicated. Therefore, this model
declares a number of features representing capabilities that not all
deployed devices support.

# I don't think that this paragraph is accurate for the YANG contained in this
draft.  The model below only defines two features, one covering IGMP Proxy and
one covering MLD Proxy.

The extensive use of feature declarations should also substantially
simplify the capability negotiation process for a vendor's IGMP / MLD
Proxy implementations.

# Again, I don't think that this paragraph is accurate and should be removed.

The YANG data model defined in this document conforms to the Network
Management Datastore Architecture (NMDA) [RFC8342]. The operational
state data is combined with the associated configuration data in the
same hierarchy [RFC8407].

## I think that this paragraph is redundant and can be removed.

The igmp-version represents version of IGMP protocol, and default value
is 2.

### represents version -> represents the version, and default -> and the default

 If the value of enable is true, it means IGMP Proxy is enabled.

## I would make this a separate paragraph.  I'm also not entirely sure why we
need this leaf, since I would expect IGMP proxy to be enabled on an interface
simply because an interface entry exists in the list.   Please can you give me
an explanation as to why it needed, and if it is needed whether it would be
better to default to true rather than false.

          leaf interface-name {
            type if:interface-ref;
            must "not( current() = /rt:routing"+
              "/rt:control-plane-protocols/pim-base:pim"+
              "/pim-base:interfaces/pim-base:interface"+
              "/pim-base:name )" {
              description
                "The upstream interface for IGMP proxy
                 must not be configured to use PIM.";
            }
            description "The upstream interface name.";

# It would be more consistent with other IETF models to just use "name" for the
interface name.  Is there a good reason to not be consistent here?  Both here,
and for downstream-interface and also for MLD.