Re: [netmod] comments on draft-schoenw-netmod-yang-types-00.txt

On Mon, Jun 23, 2008 at 10:29:47AM +0200, Martin Bjorklund wrote:
> Juergen Schoenwaelder <j.schoenwaelder@jacobs-university.de> wrote:
> > On Mon, Jun 16, 2008 at 11:31:22AM +0200, Martin Bjorklund wrote:
> > > A comment on the ipv4 address, which contains an optional zone index.
> > > Do we need to specify the expected behavior when a server does not
> > > support zone index for a particular object?
> > 
> > This might depend on the particular circumstances of the usage of this
> > object. Do you have a concrete proposal?
> 
> Maybe say that if the object does not support zones, the error-tag
> 'invalid-value' should be used.  (or maybe 'bad-element'; it's not
> clear to me when 'bad-element'/'bad-attribute' vs. 'invalid-value'
> should be used).
> 
> Can you elaborate on the particular circumstances that could affect
> the error message?

I also have trouble to understand which RFC4741 error tag is for which
purpose. Perhaps Andy can shed some light on this...

> > > RFC4001 contains some discussion about the zone index; can we refer
> > > to that rfc in this document?
> > 
> > I like to avoid a dependency on RFC 4001. Which discussion do you
> > mean? Would it be sufficient to add the following text to the
> > description clauses of ipv4-address and ipv6-address?
> > 
> >   The zone index is used to disambiguate identical address values.
> >   For link-local addresses, the zone index will typically be the
> >   interface index number or the name of an interface. If the zone
> >   index is not present, the default zone of the device will be used.
> 
> Adding this text would help.  What about the special zone index '0'
> (which in 4001 refers to the default zone)?

Do we need it? If you want the default zone, you can simply omit the
zone part. But I would also not object against having %0 mean the
same thing for those who like index numbers. But see below...

> For the ipv6 type, should we add a reference to rfc4007?

Done.

> In rfc4001, you had:
> 
>    The size of the zone index has been chosen so that it is consistent
>    with (i) the numerical zone index, defined in [RFC4007], and (ii) the
>    sin6_scope_id field of the sockaddr_in6 structure, defined in RFC
>    2553 [RFC2553].
> 
> Why do you also allow interface name now?  Can the server decides if
> interface name or index is accepted?

This is something for the WG to discuss. I allowed interface names
since this is what I believe is commonly shown on CLIs. It seems that
most people I get in touch with refer to interfaces by the equivalent
of ifName instead of ifIndex.

Sure, allowing both formats in the zone index raises a bunch of
questions that should be answered:

  - does a number take precedence to be an index number (if there is
    an interface with a matching number)

  - what are boxes supposed to generate (index numbers or names)

  - are implementations supposed to remember whether a value was
    configured as name or number.

All these questions could be avoided by just having one format, which
however requires to find agreement how NETCONF data models identify
interfaces, by name or by index number. Opinions more than welcome.

/js

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