On 7/8/25 08:40, Ben Schwartz wrote:
*From:* Paul Hoffman <[email protected]>
*Sent:* Tuesday, July 8, 2025 11:07 AM
*To:* Ben Schwartz <[email protected]>
*Cc:* [email protected] <[email protected]>
*Subject:* Re: [Ext] [DNSOP] Re: DNSOPFwd: New Version Notification
for draft-tojens-dnsop-do-not-accommodate-udp53-00.txt
On Jul 8, 2025, at 08:01, Ben Schwartz
<[email protected]> wrote:
>
> I think it could be interesting to have an informational draft that:
Interesting or useful? I don't see how this, or
draft-tojens-dnsop-do-not-accommodate-udp53 are at all useful other
than to say "haha, we're smarter than that now".
I would hope it goes without saying, but I'll say it anyway: there's no
intention to shame or otherwise deride previous standards or their
authors. The Internet and the world changes over time, and therefore so
does guidance to implementors and document authors without implying
previous approaches were less smart. I've already given my arguments on
list and in the text for why I think it's useful; reasonable minds can
disagree.
I once needed to implement a very simple DNS forwarder in an unusual
context where supporting UDP was tricky. In that case, we decided on
exactly this behavior: always return a trivial TC=1 to force DNS over
TCP. It worked fine. I’ve also occasionally needed to implement
TCP-only DNS stub behaviors, for similar reasons.
Having some documentation about this option might be helpful guidance
for future DNS implementors in situations where simplicity is more
important than performance.
Ben, I agree. I think this is a practical distillation of what I wanted
to see in the first place, just scoped to actions for implementors
rather than document authors. Would you be interested in coauthoring
such a draft? Not for 123, but after.
—Ben
_______________________________________________
DNSOP mailing list -- [email protected]
To unsubscribe send an email to [email protected]
