fujiw...@jprs.co.jp writes:

>   Using unstructured data (TXT format) is not good.

Thanks for the feedback on that.  I have wondered heavily on that
topic.  It was originally written as a text format, and we have a lot of
other cases where such text parsing exists (SPF being an example).  As
the world moves more and more to text based parsing for everything,
which I hardly say is a good thing, I wonder what the right format of
future DNS records should be.  It sounds like a good worthwhile
discussion in its own right.

In the mean time, a binary format for the record would be just fine in
my view and the record already lends itself to such a format since it's
very structured data.  It would be worth discussing once the concept in
the draft gets accepted for work by the WG.  I'm (personally) certainly
not opposed to a binary on-the-wire record format.

>   I think the multiple queries in one request is not related to DNSSEC
>   and TCP connection. They are separated elements.

We dropped the requirement for TCP in the latest version already (-03).

DNSSEC is necessary to avoid cache poisoning of illegal data (IE, to
prove you're the parent of the data as opposed to a grand parent).
-- 
Wes Hardaker
Parsons

_______________________________________________
DNSOP mailing list
DNSOP@ietf.org
https://www.ietf.org/mailman/listinfo/dnsop

Reply via email to