Re: [Cbor] Interface names (Re: changes in draft-ietf-cbor-network-addresses-05.txt)

[Brian E Carpenter <[e-mail]>]

On 27-Jul-21 21:55, Christian Amsüss wrote:
> Hello Carsten, groups,
> 
>> Interface names are local, so it doesn’t make a lot of sense to carry
>> them around between systems, which is where CBOR is mostly used these
>> days.
> 
> two applications come to mind:
> 
> * RD introspection[1] uses zone identifiers in what may be described as
>   debug output: Usually an RD won't show you a resource on a fe80::
>   address not on your link, but as an administrator you may need the
>   birds-eye view.

Yes, that is an operational use case, as also considered in RFC6874(bis)
 
> * Many network devices offer a remote way to run a ping; when the
>   process is started via CBOR, it'd be useful to give an interface
>   identifier.

How does host A know the Zone IDs (interface names) used on
host B?

(Answer: only if B has previously told A. So the use case is quite
complex.)
 
> I'm not particularly advocating that we add zone identifiers to the new
> tags (as I can use URIs or even work around completely in the former,
> and don't have a pressing need for the latter) -- but with those
> examples, a statement like "we don't specify interface identifiers
> because they shouldn't be serialized anyway" would be overreaching.

They shouldn't be on the wire except in rather special use cases.

    Brian

> 
> BR
> c
> 
> [1]: https://www.ietf.org/archive/id/draft-amsuess-core-resource-directory-extensions-05.html#name-zone-identifier-introspecti
> 
> 
> _______________________________________________
> CBOR mailing list
> [e-mail]
> https://www.ietf.org/mailman/listinfo/cbor
>