[AusNOG] "stateless TCP" for DNS

[Mattia Rossi]

On 15/11/2010 12:54, Dobbins, Roland wrote:
>
> On Nov 15, 2010, at 7:49 AM, Mattia Rossi wrote:
>
>> We also point out that DNS response sizes can be increased to 4096 Bytes, but yes, we don't mention EDNS.
>
> This is a *major* oversight which has a negative impact on the credibility of the paper; you should actually have quite a bit to say about EDNS0, including EDNS0-related breakage in middleboxes/middleblades.   It should be addressed in a revision, as you've some worthwhile questions to posit, but folks won't take you seriously if you leave such a huge conceptual and factual gap in your work.

I get your point. The current technical report is a feasibility study 
for the "stateless TCP" technique/implementation, more than an analysis 
on what the state of the current DNS is, its problems and a great 
solution to the problem. At the time of writing we were not sure about 
the situation and whether the whole idea was actually necessary, nor was 
it the focus of the report.

Thanks to the feedback so far, the ideas are getting clearer, and I see 
that we, like you, are also interested in a DNS reachability study.
Having results of such a study then, it would also be better to write a 
new, more complete paper which explains EDNS0 and the problems we're 
talking about, rather than amend the techreport.

We'll consider the options, and if you would be interested in pursuing a 
DNS reachability study with us, contact me or even better Grenville off 
list, so we can discuss it.

Cheers

Mat