>>>>Hi,
>>>>
>>>>Anyone tried the Google DNS server (8.8.8.8)?
>>>>
>>>>I just switched to it out of curiousity. I could be imagining this but it certainly seems faster than others I've used - which surprised me since I would have thought the look-up process itself wouldn't have a great overall effect on retrieving a web page.
>>>
>>>What were your ping times to your prior DNS server? I'm seeing Google about +20 msec beyond my ISP's default DNS.
>>
>>Prefacing this with the fact I may not know what I'm talking about :-{
>>
>>But isn't that just going to compare how long it takes to hit the Google DNS compared to your ISP's DNS (which I'd expect to be quicker since there would less hops) and not how long it takes for the DNS to resolve the address?
>>
>>I've no idea how you would actually get a metric on the resolution time tho'
>>
>>Regards,
>>Viv
>
>You are saying "certainly seems faster" so I was looking for something outside DNS itself that might explain the difference. I doubt Google's microseconds are any faster than others. Caching practices could be having an effect, as some sites seem to have to hit a myriad of other sites to build a page (Yahoo JS, ad netwoks, etc.). The resolution time could very well be a factor.
>
>As far as resolution times, you might look at http://www.websitepulse.com/help/tools.php and play around a bit.
>
>Or you're dreaming, Viv. :-)

OK. I played with the link you provided and the results were interesting (but obviously not definitive)

For obvious URL's (e.g. www.microsoft.com) the Google DNS outperformed my ISP's DNS.

For non-obvious one's (e.g a private URL that I maintain that probably gets no hits except from me) both DNSs intiially returned similar metrics. But the Google DNS on subsequent hits was a lot faster - presumably because it was cacheing the initial lookup?

That said , of course, the DNS part of the process was a comparatively insignificant one compared to the time taken to connect and retrieve the page. But it was interesting to note the number of times the DNS server needs to be hit to resolve page content (images etc)

Bottom line for me seems to be that Google *is* faster but the sum total of the advantage is fairly negligible.
Best,
Viv