[dns-operations] The (very) uneven distribution of DNS root servers on the Internet
Ed.Lewis at neustar.biz
Tue May 15 16:24:35 UTC 2012
At 10:12 -0400 5/15/12, McTim wrote:
>One wonders where the data came from and how up to date it is.
To me the question is the accuracy of the following statement in the
article, not the precise numbers themselves:
"One would imagine that if all things were equal, the distribution of
root servers should mirror the distribution of Internet users."
That sentence seems nice on the surface, but I don't think it is
true. As another poster notes:
"A better - although still inaccurate - metric would be root servers / ISP."
IMHO, Correlating "machine population" to "people population" is dodgy.
There's another consideration, a statistical side effect. According
to the article, Asia has 20M people per server. Oceania has 1.5M
people per server. Is it the case that Asia is under served, is it
the case that Oceania is over provisioned, or is it both, or is it
neither)? Perhaps a root server can handle 30M people's needs.
We could go on with this...on average, how many root queries are
performed while loading a web page? Using a large ISP (or any
other) recursive server, probably a number way less than 1.
In my early days, we had a person report that a user group demanded
1.544 Mbps in bandwidth needs. A bunch of us recognized the number
as something special, so we asked for a bit more info on why
precisely "1.544 Mbps." The reporter, after a few questions on his
method admitted "ok, the number is way more precise than accurate."
(As in "1.544 +/- 2" was the follow up joke.) I'd say that is the
same label that could be put on this article. The numbers probably
are precise, but don't carry the "accurate" story.
NeuStar You can leave a voice message at +1-571-434-5468
2012...time to reuse those 1984 calendars!
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the dns-operations