Alissa and Juergen: 
Thank you for catching these errors. 

Sue Hares
(Shepherd) 

-----Original Message-----
From: Juergen Schoenwaelder [mailto:[email protected]] 
Sent: Wednesday, January 18, 2017 2:09 PM
To: Alissa Cooper
Cc: The IESG; [email protected]; [email protected];
[email protected]; [email protected]
Subject: Re: [i2rs] Alissa Cooper's No Objection on
draft-ietf-i2rs-yang-l3-topology-08: (with COMMENT)

On Wed, Jan 18, 2017 at 08:01:37AM -0800, Alissa Cooper wrote:
> 
> "case interface-name {
>              leaf interface-name {
>                type string;
>                description
>                  "A name of the interface.  The name can (but does not
>                   have to) correspond to an interface reference of a
>                   containing node's interface, i.e. the path name of a
>                   corresponding interface data node on the containing
>                   node reminiscent of data type if-ref defined in
>                   RFC 7223. It should be noted that data type if-ref of
>                   RFC 7223 cannot be used directly, as this data type
>                   is used to reference an interface in a datastore of
>                   a single node in the network, not to uniquely
>                   reference interfaces across a network.";
>              }
>            }"
> 
> In RFC 7223 the data type appears to be called interface-ref, not if-ref.
> Would an example of this in this document be, say, a MAC address? 
>

Now that I read this, I must say that I do not understand the description at
all. What does 'i.e. the path name of a corresponding interface data node on
the containing node reminiscent of data type if-ref defined in RFC 7223.'
tell me? If the intention here is that it is unique across a network I think
this should be stated more clearly, that is, not as a sub-thought in a
sentence trying to explain why if-ref can't be used. And then, what is a
'network'?

And looking at the whole choice, the other options are not unique either.
For unnumbered-id, the description says "will correspond to the ifIndex
value of the interface" which very likely will clash for different nodes.
Also IP adresses may not always be unique 'across a network'. So why do we
have this "unique across a network" requirement for the interface-name case?
That is, we can't we use

/js

PS: Why is interface-name marked 'ro' in the tree diagram? It looks
    like a config true node.

-- 
Juergen Schoenwaelder           Jacobs University Bremen gGmbH
Phone: +49 421 200 3587         Campus Ring 1 | 28759 Bremen | Germany
Fax:   +49 421 200 3103         <http://www.jacobs-university.de/>

_______________________________________________
i2rs mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/i2rs

Reply via email to