>>>>> "nishant" == nishant writes:

nishant> But still i need to show that 'performance' wise BIND 9
nishant> is better than BIND 4.

My previous posting did that.

nishant> Can u please help me in deciding what kind of tests
nishant> should i really be doing to show that BIND 9 'performs'
nishant> better (or much better, as you say) than BIND 4?

Look, stop wasting your time on this pointless make-work exercise.
BIND4 is DEAD. Nobody should be running it. Consult the list archives
about this, probably from around 2000. There have been plenty of
statements about the expiration of BIND4.

Tell your users and management that BIND4 is well beyond end-of-life
and has to be replaced with BIND9. If they insist on continuing to run
dead code, get them to put that in writing and agree that they, not
you, will be held responsible for all the consequences of that action:
security holes, no support, no understanding of new DNS protocol
features, increased maintenance & administraton costs, etc, etc.

In terms of raw throughput BIND9 will always be slower than BIND4,
all other things being equal. That's because BIND9 uses more complex
data structures and is paranoid about error checking. However that
performance penalty is not noticeable in normal operations. I doubt if
anyone at your organisation would be able to tell the difference
between the response times of the two servers. Or care about it if
they did.

And as I said before, none of the root servers run BIND4. They get
thousands of queries per second -- probably at least 100 times as much
traffic as your name server. There are many good reasons why they
don't use BIND4 to squeeze out extra throughput. Not least of these is
the fact that BIND4 is dead. If BIND4 isn't good enough for the root
servers, then it's not good enough for you. Or anyone else for that
matter.