defining service ceilings

[Brad Knowles]

At 1:38 PM -0500 2005-04-11, Tim Peiffer wrote:

>  Are there any BCP documents that cover methodologies for measuring or
>  defining the maximum number of queries that a DNS server is capable
>  of?    Are there any defined test suites that we should be running?  I
>  am looking for documentation that would help validate the testing that
>  we have already done.

	Rick Jones is the acknowledged master in this field.  See 
<ftp://ftp.cup.hp.com/dist/networking/briefs/>.  Nominum also has a 
good white paper at 
<http://www.nominum.com/content/documents/CNS_WP.pdf>.  My own 
benchmarking experiences are outlined at 
<http://www.shub-internet.org/brad/papers/dnscomparison/>, although 
the RIPE44 version is more up-to-date.

>  We recently installed new hardware that replaced our authoritative
>  nameservers.  The hardware used to be run on FreeBSD 4.X, and Bind 8.X.
>  Daily  query averages are running approximately 600 qps total across all
>  of the servers.

	That's what they were seeing in terms of traffic.  Did you 
benchmark them to see how hard they could be pushed?

>                   After replacing the hardware with Sun V60 and RedHat
>  AS,  we were able to create a query source file and use the queryperf
>  utility, we were able to establish a service ceiling of approximately
>  5500qps.

	That seems rather low.  For the kind of hardware you're talking 
about, with a standard multithreading BIND9 configuration, I'd expect 
to see numbers over 10,000 queries per second, or better.

-- 
Brad Knowles, <brad at stop.mail-abuse.org>

"Those who would give up essential Liberty, to purchase a little
temporary Safety, deserve neither Liberty nor Safety."

     -- Benjamin Franklin (1706-1790), reply of the Pennsylvania
     Assembly to the Governor, November 11, 1755

   SAGE member since 1995.  See <http://www.sage.org/> for more info.