Beehive CoNoNS DNS Survey

[Kevin Darcy]

Michele Chubirka wrote:

>I don't know if this study was already addressed (didn't see anything in the
>archives), but management is asking me questions about it. Any comments?
>
>http://www.cs.cornell.edu/people/egs/beehive/dnssurvey.html
>
What about it? The survey was conducted in June of 2004, so it's a 
little dated.

There was some discussion when an indirect reference to this paper was 
posted on Slashdot recently 
(http://it.slashdot.org/article.pl?sid=06/04/26/1247240). Other than the 
idiotic "redesign DNS!" comments, a lot of the commentators pointed out 
that the survey overcounted these "dependencies" between servers in the 
resolution process.

I also noticed that they only surveyed domain names associated with 
"popular" sites, according to "Yahoo+DMOZ", rather than on a random 
cross-section. As such, some of the names of sites we consider 
mission-critical, and some we consider not-so-mission-critical ended up 
on one of their lists, but many other names of mission-critical sites, 
which rely on the same supposedly "vulnerable" infrastructure, were 
missing from the list.

- Kevin