From: Mike N. 

On Fri, 21 Jul 2006 06:23:07 -0400, "Geo"  wrote:

>It won't be any faster unless you have a really sucky dns server.

   Agreed.  I saw a review at
  http://www.wyckedone.net/?p=542
  (breathlessly glowing, so I don't give it too much weight), but it
includes a timing test.   Not enough data to be very meaningful, and he
never tested the network response time to be able to remove it from the
data.   He also didn't publish the results while connected to Charter. Even
in this case, the difference between 90 mSec and 50 mSec is negligible.

>And there
>is a downside to what they are doing, in order to have a really big cache
>they are probably setting the cache to an abnormally long time (ignoring the
>domain's TTL setting) which means it will be the last to pickup dns changes.

   I asked them, and they do not fiddle with the TTL setting.  As a bonus,
they plan to add a control to allow you to force a reload of a zone for
cases where you have made an emergency change.  A plus for those IP address
change cases you cannot plan ahead of time.

  Caching just comes from a large population of users and a large RAM
cache.  When querying an OpenDNS server with a million users, www.nls.net
is more likely to already be in cache than my local DNS server.

--- BBBS/NT v4.01 Flag-5