Re: [netmod] ietf-access-control-list@2017-10-03.yang : Can access-lists use a grouping?

On Thu, Nov 2, 2017 at 9:26 AM, M. Ranganathan <mranga@gmail.com> wrote:

> Hi Andy
>
> On Thu, Nov 2, 2017 at 11:55 AM, Andy Bierman <andy@yumaworks.com> wrote:
>
>>
>>
>> On Thu, Nov 2, 2017 at 8:34 AM, M. Ranganathan <mranga@gmail.com> wrote:
>>
>>> Hi Rob, Mahesh,
>>>
>>> Thanks for reading.
>>>
>>> On Thu, Nov 2, 2017 at 11:00 AM, Robert Wilton <rwilton@cisco.com>
>>> wrote:
>>>
>>>> Hi Ranga,
>>>>
>>>> Presumably another choice would to keep ACLs defined in one place (i.e.
>>>> no grouping required), augment with ACL model with your extra MUD + other
>>>> mgmt data, and then have a reference to that ACL from your model.
>>>>
>>>> Thanks,
>>>> Rob
>>>>
>>>
>>>  In the case of MUD ( which is just a use case driving this need ),
>>> there are local references from MUD to the ACL. MUD itself augments the ACL
>>> model.
>>>
>>> Augmentation would make (logical and design) sense if you were adding
>>> nodes that are in some way related to the ACL itself.
>>>
>>> If I wanted to Augment ACL with something that is not directly ACL
>>> relevant then Augmentation makes less sense to me from a design perspective
>>> (lets say I wanted to define a new YANG model that includes the ACL with
>>> some other system-relavant meta-data that has nothing to do with ACLs but
>>> is needed by the system in order to install an ACL).
>>>
>>> Making access-lists into a grouping and then using it in a container
>>> does not alter the ACL model as it currently stands but allows designers to
>>> use the ACL model with either augmentation or inclusion in other YANG
>>> models. Hence it improves the usability of the ACL model without altering
>>> the semantics of the current model. It is just a re-structuring but it
>>> helps the implementer.
>>>
>>>
>> Loosely coupled tables should use leafref.
>> The main concern of the NETMOD WG should be the usability of the primary
>> solution.
>>
>>
>>
>
> Not sure I understand the suggestion of using a leafref (please excuse my
> ignorance -- I am not a YANG expert by any stretch). If I used leafref,
> what leaf would I be referring to if I wanted to point to the access
> control list from another YANG model?
>

Augment is not the only way to couple data models.
You can have another list just define a foreign key (called a leafref in
YANG
since it does not have to reference a key)

>
> Also I note from the description of Access Control Lists the following
> that would indicate that it is a primary solution that one may like to
> re-use in another model.
>
>  description
>       "This is a top level container for Access Control Lists.
>        It can have one or more Access Control Lists.";
>
>
>
> If the requested change were made, would it result in excessive churn ?
>
>
I never understood why the WG wanted to change the ACL model
to its current form with containers.  Seems complicated to me.

> Thanks
>
> Regards,
>
> Ranga.
>
>
Andy

>
> --
> M. Ranganathan
>
>>
>>
>> _______________________________________________
>> netmod mailing listnetmod@ietf.orghttps://www.ietf.org/mailman/listinfo/netmod
>>
>>
>>
>>>
>>>
>>> --
>>> M. Ranganathan
>>>
>>> _______________________________________________
>>> netmod mailing list
>>> netmod@ietf.org
>>> https://www.ietf.org/mailman/listinfo/netmod
>>>
>>>
>>
>
>
> --
> M. Ranganathan
>