[dns-operations] DNS pre-resolution

[Matthew Pounsett]

I'm going to have to think on most of this.  One bit caught my eye as  
being something I can comment on immediately, though.

On 05-Dec-2009, at 17:21, David Dagon wrote:

>     I don't know if the .co authorities have seen a measurable
>     increase in traffic, with only ~2% of the browsers running
>     Chrome.  It may be hard to measure against the natural growth of
>     Internet traffic in general.  But exponential growth (a plateau
>     that cannot otherwise be explained in comparison to zone size
>     deltas) would be one thing to look for.

In my experience, query traffic growth cannot be directly related to  
zone size.  If memory serves, the first org I heard of attempting to  
make a correlation was DENIC.  I made a couple of attempts while at  
CIRA and also came up with nothing.  I recall hearing of other failed  
attempts to make a correlation, but don't recall the TLDs involved..  
someone may want to speak up.

The problem, as I see it, is that normal growth in usage of existing  
delegations so overwhelms query growth from new delegations that it  
becomes impossible to separate out the new delegation component.  I  
think to do so would require long term query analysis by QNAME, which  
is infeasible for most TLDs, and still might not produce a useful  
result.  Perhaps someone with a better statistics background than mine  
can come up with a better way to make a correlation.

Query traffic growth at the TLD does follow a very predictable curve,  
though (actually, this is more evidence to my mind against a  
correlation with registrations -- registrations don't always follow a  
predictable curve).   And that traffic curve can be very obviously  
affected by certain events on the 'net.  Small events, or slowly  
changing qualities of the 'net, can be difficult to detect; so, the  
rollout of Chrome may be lost in the noise.  But, larger or more  
sudden events can have a very clear result (the McColo shutdown had a  
dramatic effect on .ca traffic, for example).  While the Chrome  
rollout might not be detectable, the sudden appearance of a pre- 
resolving version of IE in Windows Update would be quite obvious.

Matt