database contest in c't (German IT magazine)

Hi,

in the current edition of the german computer magazine c't the results
of the database contest, which they announced last year, has finally
been published. It was based on a very simple DVD-Store web application
from Dell. [1]

Back then I spent some time modifying the original php5 application to
use PostgreSQL. I've written some stored procedures in C and optimized
the schema at obvious places. I've been short on time, though, and
couldn't do everything I wished to do. Most importantly connection
pooling didn't seem to make it into the final entry I have submitted. :-(

That might be one reason for the disappointing performance result: last
place, just before the disqualified entries. The performance trophy
(once again) has gone to the MySQL Benchmark Team... [2]

With that result I fear I did a disservice to PostgreSQL. Once again it
seems to be just slower than MySQL. I'm sorry for that. Especially
because other strengths of PostgreSQL didn't get mentioned.

OTOH I at least hacked up something. My entry was the only counting one
using PostgreSQL - which is surprising and disappointing me. (The only
other PosgreSQL entry from Alvar C.H. Freude unfortunately got
disqualified because it was sent in too late [3]). Didn't the PostgreSQL
community know about that contest? IMHO it would have been worthwhile
to start a community project and submit an entry. Just to get rid of
that 'too-slow'-image PostgreSQL still has.

At least I have learned: next time I'll try to start a community project
for such a contest!

Regards

Markus

[1]: Read more about it on http://www.ctmagazin.de/dbcontest (german,
there are english translations of the contest rules... somewhere...)

[2]: My solution (PHP5/PostgreSQL 8.1) got 120 opm (operations per
minute? Dunno anymore what that meant exactly), the best PHP5/MySQL
solution got 3664 opm. Another interesting one (IMO) was MonetDB with
1833 opm. The Java/DB2 Express one got 1537 opm, Java/Oracle 10g Express
1412 opm...

[3]: I'm quite sure Alvar's entry would have performed much better than
mine. AFAICT he has written a quite well optimized apache/mod_perl/pgsql
solution.

Markus Schiltknecht wrote:

> Hi,
>
> in the current edition of the german computer magazine c't the results
> of the database contest, which they announced last year, has finally
> been published. It was based on a very simple DVD-Store web
> application from Dell. [1]
>
> Back then I spent some time modifying the original php5 application to
> use PostgreSQL. I've written some stored procedures in C and optimized
> the schema at obvious places. I've been short on time, though, and
> couldn't do everything I wished to do. Most importantly connection
> pooling didn't seem to make it into the final entry I have submitted. :-(
>
> That might be one reason for the disappointing performance result:
> last place, just before the disqualified entries. The performance
> trophy (once again) has gone to the MySQL Benchmark Team... [2]
>
> With that result I fear I did a disservice to PostgreSQL. Once again
> it seems to be just slower than MySQL. I'm sorry for that. Especially
> because other strengths of PostgreSQL didn't get mentioned.
>
> OTOH I at least hacked up something. My entry was the only counting
> one using PostgreSQL - which is surprising and disappointing me. (The
> only other PosgreSQL entry from Alvar C.H. Freude unfortunately got
> disqualified because it was sent in too late [3]). Didn't the
> PostgreSQL community know about that contest? IMHO it would have been
> worthwhile to start a community project and submit an entry. Just to
> get rid of that 'too-slow'-image PostgreSQL still has.
>
> At least I have learned: next time I'll try to start a community
> project for such a contest!
>
> Regards
>
> Markus
>
>
>
> [1]: Read more about it on http://www.ctmagazin.de/dbcontest (german,
> there are english translations of the contest rules... somewhere...)
>
> [2]: My solution (PHP5/PostgreSQL 8.1) got 120 opm (operations per
> minute? Dunno anymore what that meant exactly), the best PHP5/MySQL
> solution got 3664 opm. Another interesting one (IMO) was MonetDB with
> 1833 opm. The Java/DB2 Express one got 1537 opm, Java/Oracle 10g
> Express 1412 opm...
>
> [3]: I'm quite sure Alvar's entry would have performed much better
> than mine. AFAICT he has written a quite well optimized
> apache/mod_perl/pgsql solution.
>
> ---------------------------(end of broadcast)---------------------------
> TIP 4: Have you searched our list archives?
>
> http://archives.postgresql.org
>

Markus Schiltknecht wrote:

Markus and the whole gang. Wow, what an opportunity to miss! In this
coming year, I will sponsor one contest winner, and I am sure others
would sponsor others. And what an idea! Let's have a contest of
submissions for such things that postgresql is better at, and mysql
worst at, which seems to be the result with this one/ michael

Markus,

> OTOH I at least hacked up something. My entry was the only counting one
> using PostgreSQL - which is surprising and disappointing me. (The only
> other PosgreSQL entry from Alvar C.H. Freude unfortunately got
> disqualified because it was sent in too late [3]). Didn't the PostgreSQL
> community know about that contest?

Apparently not. This is certainly the first I've heard of it. For one
thing, I don't speak German.

--
--Josh

Josh Berkus
PostgreSQL @ Sun
San Francisco

On Monday 12 June 2006 14:46, Markus Schiltknecht wrote:
> Hi,
>
> in the current edition of the german computer magazine c't the results
> of the database contest, which they announced last year, has finally
> been published. It was based on a very simple DVD-Store web application
> from Dell. [1]
>
> Back then I spent some time modifying the original php5 application to
> use PostgreSQL. I've written some stored procedures in C and optimized
> the schema at obvious places. I've been short on time, though, and
> couldn't do everything I wished to do. Most importantly connection
> pooling didn't seem to make it into the final entry I have submitted. :-(
>
> That might be one reason for the disappointing performance result: last
> place, just before the disqualified entries. The performance trophy
> (once again) has gone to the MySQL Benchmark Team... [2]
>
> With that result I fear I did a disservice to PostgreSQL. Once again it
> seems to be just slower than MySQL. I'm sorry for that. Especially
> because other strengths of PostgreSQL didn't get mentioned.
>
> OTOH I at least hacked up something. My entry was the only counting one
> using PostgreSQL - which is surprising and disappointing me. (The only
> other PosgreSQL entry from Alvar C.H. Freude unfortunately got
> disqualified because it was sent in too late [3]). Didn't the PostgreSQL
> community know about that contest? IMHO it would have been worthwhile
> to start a community project and submit an entry. Just to get rid of
> that 'too-slow'-image PostgreSQL still has.
>
> At least I have learned: next time I'll try to start a community project
> for such a contest!
>

If you feel strongly about it, you could still start a community project, then
get improvements and spread the word.

--
Robert Treat
Build A Brighter Lamp :: Linux Apache {middleware} PostgreSQL

Markus Schiltknecht wrote:
> Didn't the PostgreSQL community know about that contest?

It did. There were extensive discussions on the German mailing list
about it.

--
Peter Eisentraut
http://developer.postgresql.org/~petere/

On Mon, Jun 12, 2006 at 05:04:10PM -0400, Robert Treat wrote:
> On Monday 12 June 2006 14:46, Markus Schiltknecht wrote:
> > Hi,
> >
> > in the current edition of the german computer magazine c't the results
> > of the database contest, which they announced last year, has finally
> > been published. It was based on a very simple DVD-Store web application
> > from Dell. [1]
> >
> > Back then I spent some time modifying the original php5 application to
> > use PostgreSQL. I've written some stored procedures in C and optimized
> > the schema at obvious places. I've been short on time, though, and
> > couldn't do everything I wished to do. Most importantly connection
> > pooling didn't seem to make it into the final entry I have submitted. :-(
> >
> > That might be one reason for the disappointing performance result: last
> > place, just before the disqualified entries. The performance trophy
> > (once again) has gone to the MySQL Benchmark Team... [2]
> >
> > With that result I fear I did a disservice to PostgreSQL. Once again it
> > seems to be just slower than MySQL. I'm sorry for that. Especially
> > because other strengths of PostgreSQL didn't get mentioned.
> >
> > OTOH I at least hacked up something. My entry was the only counting one
> > using PostgreSQL - which is surprising and disappointing me. (The only
> > other PosgreSQL entry from Alvar C.H. Freude unfortunately got
> > disqualified because it was sent in too late [3]). Didn't the PostgreSQL
> > community know about that contest? IMHO it would have been worthwhile
> > to start a community project and submit an entry. Just to get rid of
> > that 'too-slow'-image PostgreSQL still has.
> >
> > At least I have learned: next time I'll try to start a community project
> > for such a contest!
> >
>
> If you feel strongly about it, you could still start a community project, then
> get improvements and spread the word.

Yeah, and make sure and mention it here. You'd have almost certainly
gotten volunteers, and possibly support from some of the 'PostgreSQL
companies'.
--
Jim C. Nasby, Sr. Engineering Consultant jnasby(at)pervasive(dot)com
Pervasive Software http://pervasive.com work: 512-231-6117
vcard: http://jim.nasby.net/pervasive.vcf cell: 512-569-9461

After a long battle with technology, xzilla(at)users(dot)sourceforge(dot)net (Robert Treat), an earthling, wrote:
> On Monday 12 June 2006 14:46, Markus Schiltknecht wrote:
>> Hi,
>>
>> in the current edition of the german computer magazine c't the results
>> of the database contest, which they announced last year, has finally
>> been published. It was based on a very simple DVD-Store web application
> If you feel strongly about it, you could still start a community project, then
> get improvements and spread the word.

There would further be merit to having some of this as a readily
available sample application. The MySQL guys recently relicensed a
sample database under a BSD-like license so that it could be widely
used for examples in books, presentations, and the likes.

It would be quite handy to have a decently-performant sample
application to use for examples...
--
output = reverse("moc.liamg" "@" "enworbbc")
http://linuxdatabases.info/info/spreadsheets.html
So long and thanks for all the fish.