On 24-Jul-2007, at 12:53, Akira Kato wrote:
I think it would be instructive for someone to do a measurement
exercise on a root server and identify what proportion of non-junk
queries are made with EDNS0.
I will try to analyze 48hour trace of M-Root taken in January this
year.
Couple of things:
- what is the definition of "non-junk queries"
Off the top of my head, queries that are
- well-formed
- sourced from addresses that it is possible to reply to
- are not NOTIFYs
etc.
- may we omit "junk queries" from EDNS analysis? depending on the
definition,
they are still legitimate DNS queries...
I think what is interesting to me is to examine EDNS0 availability
amongst the class of queries that can reasonably be answered. I don't
much care whether queries that can't be answered support EDNS0.
But others may be interested in different things :-)
Joe
_______________________________________________
DNSOP mailing list
DNSOP@ietf.org
https://www1.ietf.org/mailman/listinfo/dnsop
