DNS Performance Numbers

O. Hartmann ohartman at zedat.fu-berlin.de
Wed Nov 22 07:45:49 PST 2006

Mark Bucciarelli wrote:
> On Wed, Nov 22, 2006 at 11:52:47AM -0200, Marcelo Gardini do Amaral wrote:
>> The results were discussed in the following threads:
> I see the speed differences are major, but don't have a good idea
> of what 15,000 DNS queries per second means.  Is the following
> interpretation correct?
> 15,000 DNS queries per second is:
>     - 1,000 hosted domains, each getting 15 hits/second
> or
>     - 1,000 mail domains, each getting 15 deliveries/s
> or
>     - local cache for mailserver with 15,000 incoming
>       messages/s
> or
>     - some linear combination of the above.
> I guess something like SPF would adds an extra DNS query for each
> incoming message.
> I'm going to be building a DNS server soon and I'm trying to
> judge if it really matters that 6.1 is so slow.  But it's very
> hard to pass up the fourfold increase you see with NSD on 4.11
> over 6.1.
> Thanks,
> m

at the moment, your questions reveal some musunderstandings in my way of 
looking at the subject, so I should asure myself talking about the same 
problems as others do.
Means: as I see the 'numbers', they roughly show the performance of an 
untuned TCP/IP stack and/or of the speed the kernel can handle the 
'subject of being tested for'. At the conclusion, the still means Linux, 
especially their new kernel 2.6.XX outperforms both FreeBSD 4.11/UP and 
the new branch 6.1. And it was said that FreeBSD 4.11 outperforms 
FreeBSD 6.X or the new generation branch at all.
My purposes for the specific OS may be not very common and not that 
specific to DNS. Next year, in Feb, I would like to start build a number 
crunching system for my work. Due to money limitations, this would be 
done with standard SoHo or smaller server-like hardware - connection 
will be 1GbE. So, if performance of SMP capable OS and their network 
performance differ that much, a OS choice reconsideration would be 
necessary. In the past 10 years I'm now with FreeBSD, it wasn't very 
science-friendly due to some lack of suitable support by compiler 
vendors, as I wrote, i did and do not care about 10% or 15% of 
performance gain or  drainage if I could stay with the familiar OS (in 
my opinion FreeBSD was the 'better' OS in the past).

More information about the freebsd-performance mailing list