[Discuss] DNS providers

[tmetro+blu at gmail.com (Tom Metro)]

Edward Ned Harvey (blu) wrote:
> The only conclusion I'm comfortable drawing is that results vary a
> lot, and the biggest factor is network topology affecting roundtrip
> packet latency.  My results were dramatically different from those
> published on SolveDNS, and the most likely explanation is that I'm
> sitting here on a Verizon Fios network, while they tested from
> something probably different.

http://www.solvedns.com/dns-comparison/

  How do we collect the data?
  Every month, there are thousands of users that use SolveDNS as their
  DNS testing tool. This results in several million DNS queries. Each
  DNS query is timed and logged. At the end of the month, we analyze the
  data and generate a report which includes the rankings of the top DNS
  providers in terms of their DNS Query speed in milliseconds. The
  report includes the average speed, the minimum speed and the maximum
  speed across all name servers of the DNS provider. The rankings also
  include the standard deviation (SD) and the 95% confidence interval
  around the mean.

  Where do you measure from?
  Starting February 2013, we measure from Los Angeles, Dallas and New
  York. The report will display the dns speed from all the three cities
  including an average. We chose these three cities as they are
  geographically distributed across the US.


So apparently their testing tool is in the cloud (thus they control the
locations) and they do a lot of averaging across millions of requests to
end up with more stable and consistent results. They also report
deviation, so you can see how wide the swings are.

They note on the reliability report that they don't test the specified
servers on a regular interval. I'm assuming that means they just report
failures when a SolveDNS customer happens to request a test of that
server and it fails.

 -Tom

-- 
Tom Metro
The Perl Shop, Newton, MA, USA
"Predictable On-demand Perl Consulting."
http://www.theperlshop.com/