Re: Opsdir last call review of draft-ietf-rtgwg-lne-model-05

[Lou Berger <[e-mail]>]

Hi Dan,

     Sorry about the slow response.

On 1/25/2018 11:47 AM, Dan Romascanu wrote:
> Reviewer: Dan Romascanu
> Review result: Has Issues
>
> This is a very useful, well thought and well written document, which reflects
> work and discussions within the RTG and OPS areas. From an operational point of
> view it's a very useful tool in support of network operators that will manage
> and configure logical elements. I believe that the document is almost ready,
> but there are a number of issues that are worth being discussed and addressed
> before approval by the IESG.
>
> 1. The name and scope of the document as presented in the title and Abstract
> are not exactly reflecting the content. LNEs are not YANG LNEs as the title
> says, and the type of module (a YANG module) being defined is not stated in the
> Abstract.
This is fixed in -06.

> I would suggest that the document actually defines 'A Data Model and
> YANG Module for Logical Network Elements'.
okay, changed to  'YANG Module for Logical Network Elements'.

> 2. There is no reference and relationship definition in the document to the
> YANG Data Model for Hardware Management defined in
> https://tools.ietf.org/html/draft-ietf-netmod-entity-07. Actually the LNEs are
> almost similar with the 'logical entities' that were dropped from the
> netmod-entity work. It is expected that in the future network operators will
> use both data models and the respective YANG modules when managing hardware
> devices on which logical network entities are being run. Even if this
> relationship is not explicitly present in the DM, I believe that it needs to be
> looked at and mentioned in the document.
okay, it's been added to the example section as a possible model to be 
mounted.

> 3. In Section 2 I see:
>
> 'The logical-network-element module augments existing
>     interface management model by adding an identifier which is used on
>     physical interface types to identify an associated LNE.'
>
> I am wondering why the mentioning of 'physical interface types' here. What if
> the interface type in not 'physical' representing a protocol layer or sublayer
> on the device? After all, if all interfaces to be considered were 'physical' we
> could have augmented the entity hardware module rather than the interfaces
> module, as all physical interfaces are represented there as well.
excellent catch!  it should just say interfaces, it's up to an 
implementation to choose which can be bound to an LNE.

> 4. Sections 3.2 and 3.3 seem to be written for the benefit and the perspective
> of the implementations writers rather than of the operators. Are there any
> hints, advice, or indications for the operators using the module to manage
> their LNEs? These could be described also in the examples appendices, which are
> otherwise very useful to illustrate and explain the models.

I reread these sections as well as the examples and the read to me to be 
almost completely applicable/useful to both client and server - so am at 
a bit of a loss on how best to address this comment. Perhaps I'm just 
too close to the material. Can you provide a specific example of the 
type of improvement you'd like to see?

Thank you for the comments!
Lou


>
> _______________________________________________
> rtgwg mailing list
> [e-mail]
> https://www.ietf.org/mailman/listinfo/rtgwg
>