[dns-operations] Geoff Huston on DNS-over-TCP-only study.

[Geoff Huston]

On 22/08/2013, at 10:32 AM, David Conrad <drc at virtualized.org> wrote:

> Geoff,
> 
> I personally think this is really interesting work. A question about methodology:
> 
> On Aug 21, 2013, at 4:36 PM, Geoff Huston <gih at apnic.net> wrote:
>> - Our experiment used a modified DNS server that truncated all UDP at 512 bytes, and over 10 days we enlisted some 2 million end clients to perform a set of tests by using online ads. The ad used a very wide geographic and network variety, so there is good grounds to see this set as a reasonable representative sample of the internet's end user population.
> 
> If I recall correctly, you're using a Flash thingie to do this.  Is that right?
> 
> If so, have you looked at how platforms that don't do Flash (notably, Apple IOS-based devices by default) behave (at least in a lab)?  I know those devices had an ... interesting impact on the IANA servers providing the root trust anchor... 

We have used flash becuase flash is used by Google Ads and we use Google's ad distribution network to feed the ads. I've seen work by a research crew at UCSD who mounted an ad using iframes and javascript, but their usenix paper did not name their ad distribution network, so I am trying to see if we can target the  non-flash platforms (i.e. Apple i* devices) using a different ad network.

Parenthetically, I see my vanilla Mac (OSX 10.8.4) does not use extended UDP sizes in its queries to the local resolver, so it needs these local resolvers to pass back large queries using TCP.

Geoff