Re: [netmod] Fw: New Version Notification for draft-bjorklund-netmod-rfc7277bis-00.txt

[Martin Bjorklund <[e-mail]>]

Hi,

Thanks for reviewing this document!


Robert Wilton <[e-mail]> wrote:
> Hi Martin,
> 
> I have reviewed this document (mainly as a diff against the 7277).
> 
> In summary, the transformation to NMDA looks good.   Given that it is
> a fairly formulaic conversion, I hope that the WG will be able to
> quickly move to WG adoption, and even WG last call!
> 
> 
> I did have a few minor comments (and spotted one bug):
> 
> 1. leaf "origin" under IPv4 neighbors should be marked as "config
> false".

Fixed.

> 2. Sometimes the YANG module description refers to "the intended
> configuration datastore" or "the operational state datastore", but I
> think that it would be better to just refer to "intended
> configuration" and "operational state".

Yes I agree, and I actually inteded to use the term "intended
configuration" and "operational state", but missed that.

The terminology should be aligned in all our documents, so I'll make
sure to use this terminology also in the hardware (entity) draft, and
the interfaces draft.

>   I.e. I think that it is OK
> for the draft to reference the different datastores, but I think that
> it might be better if the YANG modules don't.  This is partly in the
> sense that the YANG modules are just schema for configuration and
> state data, and that I see the use of datastores is effectively an
> IETF decision on how that configuration/state is instantiated.
> 
> To take the analogy further, I think that the it is conceivable that
> the source OpenConfig YANG models could also be usefully structured in
> the same NMDA style.  This would allow the OpenConfig models to work
> better with NETCONF/RESTCONF implementations that support NMDA.  An
> alternative implementation choice would be to use a script to convert
> NMDA style OpenConfig YANG models to equivalent existing OpenConfig
> style with their split config/state containers.
> 
> A second justification (if you don't like the one above), is that the
> configurable leaves could also be set via a dynamic configuration
> datastore such as via I2RS, hence it wouldn't necessarily always be
> via the intended datastore.
> 
> 
> 3. I think that it might be useful to add a comment to the IPv6
> link-layer-address leaf description to indicate that it wouldn't
> expect to be populated if the associated "state" is set to
> "incomplete".

Ok, done.


/martin


> 
> 
> Thanks,
> Rob
> 
> 
> On 21/08/2017 13:06, Martin Bjorklund wrote:
> > Hi,
> >
> > This document defines an NMDA-compliant update to the IP model
> > (RFC 7277).
> >
> > I would like to ask the WG to adopt this individual draft as a working
> > group document.
> >
> >
> > /martin
> >
> >
> >
> > _______________________________________________
> > netmod mailing list
> > [e-mail]
> > https://www.ietf.org/mailman/listinfo/netmod
>