RE: [PHP] Independant ASP & MSSQL Vs PHP MySQL Tests

Okay, here I go again :) ODBC is not inherently slower, and is in fact faster than native drivers if written correctly. Anyone from Easysoft want to back me up here? Michael: Granted, I wouldn't use Microsoft's ODBC on Windows and native drivers on linux, but there are alternatives. Andy: to level the playing field, why don't you use Openlink's drivers on both platforms, via ODBC? This would also allow rapid testing of other databases and platforms, and since our drivers are free at 10 connections or less, there isn't a commercial consideration, and I would be happy in assisting with any setup. Best regards, Andrew ---------------------------------------------------- Andrew Hill Professional Services Consultant OpenLink Software http://www.openlinksw.com Universal Database Connectivity Technology Providers -----Original Message----- From: Michael Kimsal [mailto:michael@tapinternet.com] Sent: Wednesday, August 02, 2000 11:04 AM To: php-general@lists.php.net Subject: Re: [PHP] Independant ASP & MSSQL Vs PHP MySQL Tests Hello... First off, as a moderately active ASP programmer, I must jump in here a bit. Don't use ODBC on NT. No one will seriously trust those numbers, as ASP under NT connecting to SQL server (7.0) using OLEDB to connect has been a recommended method for about a year. connection.Open "Provider=SQLOLEDB; Data Source=machine; User ID=ID; Password=PASSWORD" It is much faster than regular ODBC. Anyone inspecting your numbers would cry 'foul' that you used ODBC on NT but native drivers under PHP. Why test PHP on W2000? The docs themselves say it's not ready for deployment on production machines (at least my readme.txt said that). Also, it's not optimized for NT. You're setting yourself up for disappointing numbers. In our experience NOTHING will beat optimized ASP/VBScript/SQL7 under Windows. I wrote that wrong. No other similar platform (PHP, CF, Perl) will beat optimized ASP on Windows. MS *WROTE* the bloody thing - they know all the nooks and crannies and hidden memory stuff. You can't beat the enemy on their own ground. They've got the 'home field advantage'. Choose another similar catchphrase if you don't like those. I'm not sure what benefit 'vanilla' installs will be in judging things. You probably have seen many reactions to NT/Linux comparisons where people always say 'but they didn't do such-and-such tweak to Linux - then it would have kicked butt!'. Take into account as many enhancements on each side that you can - most professionals in the NT side of things will tweak to make their code as fast as possible - do the same for PHP. ASP has the getrows function, which many people use - there isn't an equivalent under PHP (see archives for my question on this). Be prepared to answer questions from prospective clients about why you didn't compare this functionality (or use stored procs as well). There are many many enhancements to 'vanilla' NT that most ASP people know something about. Be prepared to accomodate for these discrepancies. I'm going thru the same issues now with a client - we're probably migrating from ASP/NT to Linux/PHP. I've told them that straight code for code, the Linux might be a bit slower. That's not the whole point - you can always throw a faster processor or more boxes at the problem. The better argument, imo, is code and maintenance. A remote NT machine is harder to administer than a remote Linux machine. PHP code is, imo, handier and bit more elegant than VBScript. Linux/PHP/MySQL may indeed be faster, but MySQL also doesn't have all the features of SQL7. That's a benefit and hindrance. People familiar with MSSQL will feel cramped rewriting things in PHP or MySQL that are easier in MSSQL. But MySQL can be much faster because it doesn't have the overhead associated with MSSQL. Keep that in mind. We did similar tests a few months back - results are up at http://www.tapinternet.com/bench.php3 Andy Woolley wrote: > Henrik, > > Good point. > > I will be taking the easiest option with regard to development. As this will > be what the majority of our fellow developers will be using most of the > time. ASP will have to use ODBC to connect to either SQL Server or MySQL. > This is the way Microsoft tell you to connect and that is the way I will be > connecting. PHP on the other hand contains native MySQL functions that are > easier (in my opinion) to use than ODBC. > > The general trend of the tests will be Vanilla installs and the easiest to > develop options. I'm not as fluent with ASP as I used to be because almost > everything I do now is PHP/MySQL, so I wouldnt know what other db connection > alternatives have been developed recently for the MS platform and I > appologise for that. > > Andy. > > ----- Original Message ----- > From: "Henrik Johansson" <henrik.johansson@radius.se> > To: "'Andy Woolley'" <andy@reftech.co.uk> > Cc: <php-general@lists.php.net> > Sent: Wednesday, August 02, 2000 12:02 PM > Subject: RE: [PHP] Independant ASP & MSSQL Vs PHP MySQL Tests > > > Hi! > > > > It would be intresting to see some benchmarking/stability-tests with > > ASP/PHP4 using > > the ODBC/JDBC unified database-interface against > > MSSQL/MYSQL/Oracle/PostgresSQL > > as well! :-) > > On the other hand, I guess the wishes for intresting combinations could go > > on for ever so > > maybe best not taking my cursiosity too seriously. ;_) > > > > Best Regards, > > Henrik J. > > > > > ---------- > > > From: Andy Woolley[SMTP:andy@reftech.co.uk] > > > Sent: den 2 augusti 2000 11:32 > > > To: php-general@lists.php.net > > > Subject: [PHP] Independant ASP & MSSQL Vs PHP MySQL Tests > > > > > > Greetings, > > > > > > Over the next week or so I am going to run some performance tests on the > > > following: > > > > > > Microsoft 2000 Server running IIS 5, ASP and SQL Server > > > Microsoft 2000 Server running IIS 5, PHP4 and SQL Server > > > Microsoft 2000 Server running IIS 5, PHP4 and MySQL > > > > > > I am then going to Compare the results from the above with FreeBSD, PHP4 > > > and MySQL. > > > > > > I will create a Database Table with around 10,000 records and then test > > > Insert, Update, Delete and ofcourse Select statements. > > > > > > The results will be posted on the web and I'll issue the address as soon > > > as it's set up. > > > > > > What I need from you guys is some feedback on what you would like > testing. > > > > > > > > > I am going to run the tests so that I can have some statistics that I > can > > > pass to business partners when they question our use of Open Source > > > products as opposed to Microsoft's. i am also doing it for my own > > > curiosity. This will be useful for everyone, I hope. > > > > > > So, let me know what you think. > > > > > > Regards > > > > > > Andy Woolley. > > > > > > > -- > PHP General Mailing List (http://www.php.net/) > To unsubscribe, e-mail: php-general-unsubscribe@lists.php.net > For additional commands, e-mail: php-general-help@lists.php.net > To contact the list administrators, e-mail: php-list-admin@lists.php.net -- ========================== Michael Kimsal http://www.tapinternet.com 734-480-9961 -- PHP General Mailing List (http://www.php.net/) To unsubscribe, e-mail: php-general-unsubscribe@lists.php.net For additional commands, e-mail: php-general-help@lists.php.net To contact the list administrators, e-mail: php-list-admin@lists.php.net