“Finding solid performance data to help choose among competing technologies is as tough as creating the data in the first place. This is particularly true in the database space, where database vendors routinely use no-benchmarking clauses in their license agreements to block publication of benchmarks of which they do not approve. […] We tested IBM’s DB2 7.2 with FixPack 5, Microsoft Corp.’s SQL Server 2000 Enterprise Edition with Service Pack 2, MySQL AB’s MySQL 4.0.1 Max, Oracle Corp.’s Oracle9i Enterprise Edition 9.0.1.1.1 and Sybase Inc.’s ASE (Adative Server Enterprise) 12.5.0.1.” Read the rest of the benchmark article at eWeek. Our Take: Hey, where are my PostgreSQL benchmarks?
Server Databases Clash – eWeek Benchmarks
About The Author
Eugenia Loli
Ex-programmer, ex-editor in chief at OSNews.com, now a visual artist/filmmaker.
Follow me on Twitter @EugeniaLoli
Go mysql!!, it will only get better!
What is wrong with these people, dont they realize Post can easily match Oracle in response time and scale??? They have a very good JDBC driver as well. I have it running an a BSD box with a Dual XEON system. Runs smoothly on heavy load.
Would have like them to include this Database as well.
This tests the entire stack (app server, web server, etc.).
When comparing the seperate .NET results with the other results the only thing we can conclude from this article that the SQL_Server/.NET combo easily outerperforms pick_any_DB/J2EE combo.
Something we allready knew.
on which the databases where running.
This is retarded. The JDBC driver could easily be the most significant factor in all of these differences. While, in my experience, I’ve gotten best java->sql perf from MySQL/mm.mysql.org-2.0.7-bin.jar overall, I’ve noted HUGE differences in perf when using, say MSSQL7/BEA Kona vs. MSSQL7/j-netdirect (j-netdirect is way faster in my tests).
So, as much as they acknowledge “MSSQL’s JDBC driver had some issues”, if they tested multiple JDBC drivers/RDBMS, they’d probably see that that might be more significant than anything. One of the big reasons I use MySQL personally (I have licenses for Oracle and MSSQL2000) is that the MM’s jdbc driver is very very good and free. It is also under rapid, active development so it should continue to improve.
I can’t believe they didn’t test PostgreSQL!
RedHat PostgreSQL rules!
ciao
yc
I don’t believe that IBM DB2 benchmark at all.
This drop around 600 users is so unnatural.
I would expect to see some kind of logarithmic curve not
the bell shaped one. Most likely it’s due to misconfigured server.
Well, just another proof of old adage: there is no worse lie than statistics and bechmark is the essence of it.
I guess when there is no news in OS world we’ll get news from DB world. May be we’ll get some reviews of mail clients
and mp3 players one day.
They missed out Informix in their database list as well.
They also should have looked at the performance under another OS other than Windows. Only SQL Server works well under windows.
Certainly my company has always achieved huge speed benifits when we switch to real hardware under a real OS.
Having a genuine, standard runtime environment for algorithmic object-pooling, connection-pooling and thread-pooling (COM+) gives Windows a huge efficiency advantage in distributed apps over UNIX. UNIX-land has no STANDARD runtime environment for coders to write to. That’s why a 32-bit multiprocessor machine like the Unisys ES7000 can eat bigger Sun and IBM 64-bit multiprocessor boxes for lunch when running SAP. See…
http://www.unisys.es/news/releases/2001/oct/10228070.asp
The UNIX ‘solution’ to every application performance problem is to just throw more hardware at it and hope that it goes away (yawn).
Sun will be dead as soon as IA-64 starts gaining traction in the server space, they just don’t realize it yet.
Agreed that MySQL is the enviable David amongst a number of Goliaths!! But for the features your sacrifice, the performance for a free database cant be beat. I agree with Linux_baby, as time goes on, it will only get better, since it is not really under any big pressure to compete for business as are the Oracles, Sybases and SQL Servers!!
http://www.freebsdforums.org
As Ewoud Jansen pointed out, these benchmarks are affected not only by the performances of the database but also of the tcp/ip stack, web server and all the in-betweens.
A proper benchmarking test of the DB itself would directly perform the transactions with the server with a fixed given data-set; repeating the process for various configurations and data access patterns.
I also think the benchmarks of .NET where MSSQL is the only competitor was kinda lame, what was he trying to prove?
It is a poor benchmark of the Nile app with various combinations of components. This kind of benchmarks is properly done by the Transaction Processing Council. This non-profit organization is sponsored and monitored by all major vendors.
Check it out at http://www.tpc.org
I really think this test is not complete. I wonder why they use a Window machine as database server? They could also have used a nice Sun machine or a HP-UX or even a Linux machine for database server. That would make it possible to test PostgreSQL, Oracle (again because it’s also available on Linux) en MySQL and they could include Berkeley DB…
In my experience MySQL has a rather incompatible SQL-syntax. String functions are more C-like than SQL. Postrgress seems to follow the standard more closely and allow easier switching between diffrent dbs.
Of course I don’t know about the latest MySQL releases.
eurgh! what a crock. mysql vs db2/oracle? ouch. this is such a joke. and i dont buy the db2 result either! i use db2 every day (albeit on AS400’s, not crap pc’s).
all these db’s are apples vs oranges.
low end – mysql
mid range – postgress, sql server/ase
high end – db2, oracle
We use MySQL at work and on our web site, it basically
blazes for our application, and you can’t beat the price.
We were siked to see such good benchmarks in Version 4.
Dano.
Postgresql 7.2 is a standard SQL database that positively rocks on the PC, as long as you use something like BeOS with SMP!