Re: updated dname draft-08

[Paul Vixie <paul@vix.com>]

> | so if a presented DNAME is superior to a presented CNAME, then a receiver
> | can assume that they are related (CNAME was synthesized)?  what if there
> | is more than one label of difference between them, and one of those labels
> | has other data, and there is no actual relationship, and the CNAME is
> | real?
>  
> If there is more than one label of difference, the DNAME applies.  Those
> other labels are not allowed to have data (no data below DNAME).  If the
> CNAME is real, then there is some sort of data below a DNAME, which is not
> allowed. The DNAME applies to all names ending in that suffix, also names
> with lots of labels more.

there's a lot of subtlety there.  but i agree, it works.

> | i think you need a signalling method to say, the CNAME was generated, if
> | you want initiators to be able to reliably ignore them when a DNAME is
> | also presented.
> 
> Is the current UD bit copied into the response, if older servers won't
> copy it into the response, good enough?

it wouldn't, since older authority servers also synthesize, but according to
what you pointed out above, it doesn't matter in that case.

--
to unsubscribe send a message to namedroppers-request@ops.ietf.org with
the word 'unsubscribe' in a single line as the message text body.
archive: <http://ops.ietf.org/lists/namedroppers/>