[Phplogcon-dev] Database abstraction layer for phpLogCon

Rainer Gerhards rgerhards at hq.adiscon.com
Wed Dec 7 15:48:09 CET 2005


Well... I've gone through the references and my guess is that Pear will
probably be not that bad in our case (though ADOdb might be something we
should look at). My reason is that I think we do relatively simply
queries. Anyhow, these simple queries can relate to a lot of i/o at the
database itself, which probably turns out to be the botleneck. Of
course, nothing of this is verified, but I have the strong impression
that performance will not be that much of an issue (well, to be
precisely "performance of the abstraction layer" - performance per se
*is* an issue, especially with the potentially huge amounts of data we
have in syslog... ;)).

So my educated (but unverified) opinion is that it would probably be
worth looking at Pear. I Am still of the view that native DB support via
our own layer is causing more trouble than it is worth.

My 2cts...

Rainer

> -----Original Message-----
> From: phplogcon-dev-bounces at lists.adiscon.com 
> [mailto:phplogcon-dev-bounces at lists.adiscon.com] On Behalf Of 
> Michael Meckelein
> Sent: Wednesday, December 07, 2005 12:39 PM
> To: phplogcon-dev at lists.adiscon.com
> Subject: Re: [Phplogcon-dev] Database abstraction layer for phpLogCon
> 
> > Of course, the question is what implications Pear has - e.g.
> > performance-wise.
> 
> Of course, abstraction layer have naturally impact on 
> performance. I did
> some research about pear::db performance. I was surprised some say
> "PEAR::DB code will run at about 3/8 the speed of the equivalent
> DBMS-specific code" [1]. Also found some benchmark indicating that is
> true [2][3].
> 
> As I already wrote, I have used pear::db in small projects 
> and it works
> great. The impact of the abstraction layer was hardly noticeable
> (subjective), but I have not made any performance testing. 
> 
> "Is PEAR DB worth using?" [4] found at comp.lang.php. It's a good
> decision indicates that it worth to use a db abstraction layer. I
> support this approach since I know about the trouble, testing and time
> effort for developing your own db wrapper.
> 
> Michael
> 
> [1] Impaired performance of pear::db
> http://www.hudzilla.org/phpbook/read.php/9_6_4
> [2] simple benchmark (08/13/02) comparing some db abstraction layer
> http://freshmeat.net/screenshots/30313/
> [3] Comparing ADODB with PEAR DB, MDB, dbx, Metabase and Native MySQL
> http://phplens.com/lens/adodb/
> [4] Is PEAR DB worth using?"
> http://groups.google.com/group/comp.lang.php/browse_frm/thread
> /1d1dca65e
> 2c097af/f3d8197af6c9d4f8?lnk=st&q=pear+db+performance&rnum=2&hl=en
> 
> _______________________________________________
> Phplogcon-dev mailing list
> http://lists.adiscon.net/mailman/listinfo/phplogcon-dev
> 



More information about the Phplogcon-dev mailing list