I realize that the anycast discussion was meant by Karl as an example.
But there was precisely one technical concern I had when discussion got
going. And that was that if something went wrong- meaning that someone
was returning bad data- the IP address wouldn't necessarily provide a
clear answer as to who the source of the bad data is.
I expressed this concern privately to Paul Vixie who provided me a very
satisfactory answer: you can query the name server for a record that
will provide you uniquely identifying information. I'll let Paul
describe this, but it amounts to the borrowing of an unused class for
management purposes.
While there is always room for improvement of course, Paul's answers
make it clear to me that the root folk have given this some fairly
careful thought. I also agree with Paul on another point- different
methods used by different servers ARE a good thing, so that no one
logical attack could take them all out.
Good documentation is also really important. It turns out there is some
for F, at least. See http://www.isc.org/tn/isc-tn-2003-1.html by Joe Abley.
Eliot