Re: [i2rs] Yangdoctors early review of draft-ietf-i2rs-yang-l3-topology-10

[Ladislav Lhotka <[e-mail]>]

Hi Alex,

Alexander Clemm <[e-mail]> writes:

> Hello Lada,
>
> Thank you very much for your detailed review.   Will plan to update the
> draft over the coming two weeks or so (as I may be taking a few days off).
> Most of the comments should be reasonably straightforward to accommodate.
> However, I do have comments regarding #1 and #4.  
>
> On #1 (presence containers vs identities).  Unless you feel very strongly
> about this, we would prefer to keep this as is, not revert to identities.
> This node is subject to quite a few "when" condition checks (when it comes
> to topology-specific augmentations) which are very simple this way.

It means that the type info can change at run time, right? Then it
probably needs some more explanation and examples because at this level
the use cases are not very clear.

> Changing this makes the conditions more complex and has ripple effects
> through dependant modules.  Note that a topology can have multiple types; a
> single leaf with identity would not be sufficient, it would need to be a
> leaf-list.  On #4, (URIs) we had multiple discussions in the past but have
> always come back to that; I would prefer to not change this at this point.
> Please note that the model is implemented and deployed today, e.g. in Open
> Daylight.

Does OpenDaylight UI expose human users to these URIs?

Lada

>  
> Thanks
> --- Alex
>
> -----Original Message-----
> From: i2rs [mailto:[e-mail]] On Behalf Of Susan Hares
> Sent: Wednesday, July 26, 2017 5:40 AM
> To: 'Ladislav Lhotka' <[e-mail]>; [e-mail]
> Cc: [e-mail]; [e-mail];
> [e-mail]
> Subject: Re: [i2rs] Yangdoctors early review of
> draft-ietf-i2rs-yang-l3-topology-10
>
> Lada:
>
> Thank you for reviewing the draft with such a careful eye to now and the
> future.  Please look for a response from Alex within a few days, .
>
> Sue 
>
> -----Original Message-----
> From: Ladislav Lhotka [mailto:[e-mail]] 
> Sent: Wednesday, July 26, 2017 4:46 AM
> To: [e-mail]
> Cc: [e-mail]; [e-mail];
> [e-mail]
> Subject: Yangdoctors early review of draft-ietf-i2rs-yang-l3-topology-10
>
> Reviewer: Ladislav Lhotka
> Review result: Ready with Issues
>
> This YANG module and I-D belong to an extensible suite of data models
> addressing multi-layered network topology. It is an interesting, even if
> somewhat atypical, application of YANG.
>
> The entire network topology suite seems well thought out, and it is apparent
> that the document has already undergone several rounds of discussions and
> iterations. It is also positive that the document deals with possible
> overlaps with other YANG modules such as ietf-interfaces or ietf-hardware.
> Of course, an ultimate acid test of the network topology suite will come
> with applications to real-life network.
>
> My review below raises several issues that need to be addressed (especially
> #1). Some of them are related to design decisions described in
> draft-ietf-i2rs-yang-network-topo. Apart from that, I believe the document
> is ready to be published.
>
> *** Comments to draft-ietf-i2rs-yang-network-topo-14:
>
> 1. With YANG 1.1, the network type information looks like a perfect
>    candidate for identities (with multiple inheritance). Instead, it
>    is modelled with presence containers, which is cumbersome and
>    redundant. I don't see any reasons for doing so, if there are any,
>    then they should be explained in sec. 4.4.8.
>
> 2. The document defines "supporting network" and then says "A
>    supporting network is in effect an underlay network." In the
>    subsequent text "underlay network" is used almost exclusively. So
>    perhaps the term "supporting network" should be dropped altogether?
>
> 3. The text should be better aligned with the terminology of
>    draft-ietf-netmod-revised-datastores. In particular,
>    "system-controlled" should be used instead of "server-provided".
>
> 4. Is it necessary to use URIs for identifying all objects in the
>    data model. URIs are difficult to use, so applications are likely
>    to introduce some abbreviations and keep an internal mapping, which
>    could cause problems, e.g. when matching objects in notifications
>    with a GUI representation. In my view, it should be sufficient to
>    use URIs for network-id and plain strings for other IDs, because
>    all other objects are encapsulated inside a network, so their name
>    conflicts shouldn't matter.
>
> *** Comments to draft-ietf-i2rs-yang-l3-topology-10
>
> 5. The type of "router-id" should be "yang:dotted-quad" and not
>    "inet:ip-address" because the latter means both IPv4 and IPv6
>    address, possibly also with a zone index.
>
> 6. Both prefix and link attributes include the "metric" parameter. It
>    should be explained what they mean. In the context of ietf-routing
>    the option of including "metric" as a general route parameter was
>    discussed and finally rejected because different routing protocol use
>    metrics with different semantics and properties. I wonder if it is
>    the same case here. 
>
> *** Formal issues and typos
>
> 7. Both documents should refer to draft-ietf-netmod-yang-tree-diagrams
>    rather than describe the notation of tree diagrams in the text.
>
> 8. Sec. 7 in draft-ietf-i2rs-yang-l3-topology-10: s/moodel/model/
>
>
>
> _______________________________________________
> i2rs mailing list
> [e-mail]
> https://www.ietf.org/mailman/listinfo/i2rs
>

-- 
Ladislav Lhotka
Head, CZ.NIC Labs
PGP Key ID: 0xB8F92B08A9F76C67