Let's see, millions of full-service resolvers, times the packet-count
differential between UDP and TCP, times the average reload/restart frequency of
those full-service resolvers per day/week/month. Can't a case be made from
sheer volume?
Sorry for bringing math into the discussion.
- Kevin
-----Original Message-----
From: DNSOP [mailto:[email protected]] On Behalf Of ????
Sent: Friday, October 16, 2015 12:49 PM
To: Paul Vixie
Cc: Shane Kerr; dnsop WG; Paul Hoffman
Subject: Re: [DNSOP] Closing out issues in draft-ietf-dnsop-resolver-priming
At Fri, 16 Oct 2015 08:35:30 -0700,
Paul Vixie <[email protected]> wrote:
> > I have separate issue, which is this text:
> >
> > The priming query MUST be sent over UDP (section 6.1.3.2 of
> > [RFC1123]).
> >
> > This seems like a super-strong recommendation that doesn't actually
> > help anything in operation. Further, it seems to conflict with a
> > general desire to make TCP the equal of UDP in DNS.
>
> i do not share, or approve of, any such "general desire". i think the
> text is correct as written.
Although the existence of the "general desire" might be debatable, I think
Shane has a valid point. The requirement level of the TCP support has been
already tightened in RFC5966 (from a SHOULD in
RFC1123 to REQUIRED), so it doesn't make much sense to me to use a MUST in this
document referring to the older RFC. At the very least, if we want to keep the
MUST, there should be a different reason than the reference to RFC1123.
--
JINMEI, Tatuya
_______________________________________________
DNSOP mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/dnsop
_______________________________________________
DNSOP mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/dnsop
