Re: [PATCH] dnscache: merge similar outgoing udp packets
Krzysztof Gajdemski <[email protected]>
| Newsgroups | gmane.network.djbdns |
|---|---|
| Organization | Dragon Globulka Z |
| Message-ID | <[email protected]> |
24.04.2009, 09:15:09, Jeff King wrote:
> On Thu, Apr 23, 2009 at 10:48:00AM +0200, Krzysztof Gajdemski wrote:
> > I've just tested Your patch in rather loaded dnscache cluster. Patched
> > dnscache was installed on only one server within the cluster so
> > performance between patched and unpatched version could be easily
> > compared. During a period of testing (about one hour) I had no visible
> > name resolving problems. Query merging engine also worked just fine.
> > Unfortunately, there was a serious performance issue.
> I am not too surprised by that. The data structures are very naive, so
> for a heavily loaded dnscache, you end up with some O(MAXUDP^2)
> behavior. I'll try to take some measurements, and then replace the
> main linear search for duplicate matching with something with better
> algorithmic complexity.
Yes, this explains the problem pretty well. Thanks for the answer.
If you'll be ready to provide any optimized/tuned code (even in alpha stage),
I'll be happy to do similar performance tests in my environment and
share the results.
Regards,
k.
--
Krzysztof Gajdemski | songo (at) debian.org.pl | KG4751-RIPE
Registered Linux User #133457 | BLUG Registered Member #0005
PGP key at: http://s.debian.org.pl/gpg/gpgkey * ID: 3C38979D
SzanujÄ was wszystkich, ktĂłrzy pozostajecie w cieniu - Snerg