Impact of having loghost defined in DNS...?

Jonathan de Boyne Pollard J.deBoynePollard at Tesco.NET
Fri Dec 10 04:41:11 UTC 2004

DC> I can't find the nifty pie chart I saw the other day that showed
DC> something like 2% "real" queries.

If it's the one from the old CAIDA study results announcement, note that 
the headline conclusions derived from that study are utter bunkum.  
Based upon the graph published in the announcement, rather than 98% of 
the queries being unnecessary (as the headlines state), only a mere 11% 
(a far less headline-grabbing number) of the queries are unequivocally 
"unnecessary".  19% of the queries are "real" ones, according to the 
graph; and indeed anywhere up to 89% of the queries could be "real" 
ones.  (The published CAIDA data simply aren't sufficient for 
determining the precise figure.)

The CAIDA headlines were debunked almost two years ago.  Please stop 
repeating them.


Be highly suspicious of claims made by CAIDA in this field.  It relies 
upon second-hand and erroneous data, rather than finding out for itself, 
and comes up with wild and outlandish, headline-grabbing, hypotheses and 
conclusions, completely overlooking many glaringly obvious, and far more 
prosaic, possibilities.


More information about the bind-users mailing list