Re: [DNSOP] Priming query transport selection

[Jim Reid <[e-mail]>]

On 13 Jan 2010, at 20:49, Alex Bligh wrote:

> Current operational practice would result in DO clear packets  
> fitting within 4096 bytes, so no need for TCP when DO is clear.

I don't think that's always the case Alex. See the lengthy discussion  
in this list about datagram fragmentation and broken middleware boxes  
that don't grok EDNS0. [Or do EDNS0 with a 512 byte buffer size.  
Sigh.] Mind you, some of those boxes will also barf on TCP DNS traffic.

> I thought it might be an error as the section title in your
> proposed text doesn't match the section text, but that's
> actually because two section titles are the same which
> I think is the problem.

Indeed: a stupid cut&paste error. Sorry.

> Thinking about it, a total prohibition (at MUST level) of using TCP
> is probably a bit harsh given we don't even know they have UDP
> connectivity. Perhaps "MUST issue the priming query *first*
> over UDP", or use a SHOULD.

SHOULD is more appropriate than a MUST IMO. If the resolver has a  
priori knowledge that UDP/EDNS0 will fail for some reason, forcing  
them to do that and then revert to TCP or whatever would be a Bad Thing.

The preferred approach might probably be along these lines:
	[1] EDNS0 + DO with a buffer of 5-8K (ish)
	[2] TCP + DO when [1] fails
	[3] EDNS0 + DO + 1.5K (ish) buffer if [2] fails
	[4] EDNS0 (no DO) with a 1.5K (ish) buffer
	[5] Vanilla UDP (no EDNS0) if [4] fails

I think it would be helpful if the guidance on priming queries was  
split into 3 categories: resolvers that speak DNSSEC, those that are  
not DNSSEC-aware but speak EDNS0 and resolvers that are ignorant of  
both protocols. They'd start at [1], [4] and [5] respectively in the  
scenario above. The optimal priming behaviour for each may well be  
different, particularly wrt EDNS0 buffer minima and maxima. It would  
be good to give an explanation for those buffer sizes too in case  
we've all forgotten about that when revisiting the issue 5-10 years  
from now.

Perhaps the recommended resolver behaviour should apply to all queries  
and not just the priming query?