On 13Nov 2007, at 3:43 PM, Paul Vixie wrote:
this thread appears to have ratholed. bert proposed language. are we there yet?
You mean: "Implementations MUST NOT use Query-IDs that can easily be predicted"Speaking for myself: I found the discussion helpful to understand the issue in depth. Now I am comfortable with the above sentence. And for what its worth, the recursive nameserver of which development I am sideways involved in follows the behavior you describes too :-)
--Olaf ----------------------------------------------------------- Olaf M. Kolkman NLnet Labs http://www.nlnetlabs.nl/
Attachment:
PGP.sig
Description: This is a digitally signed message part