Looks like with Oracle’s recent announcement all of the big names in the commercial database market have free versions that can be used in production environments. There is Sybase ASE Express Edition for Linux, Microsoft’s SQL Server 2005 Express Edition, and Oracle Database 10g Express Edition.

Here’s a short feature comparison on the very basic stats.

Oracle Database 10g Express Edition
4 GB of data
1 CPU
1 GB of RAM
One instance per server
Downloads For Windows or Linux

Microsoft’s SQL Server 2005 Express Edition
4 GB of data
1 CPU
1 GB of RAM
Windows Only

Sybase ASE Express Edition for Linux
5 GB of data
1 CPU
2 GB of RAM
Linux

Each of these databases are exceptional pieces of technology. I have worked mainly with Oracle and SQL Server and am really blown away with the features that have been packed into Oracle 10g and SQL Server 2005. I think these moves by the commercial database companies to release free versions of their software could really have an impact on the open source databases typically used to build small to medium sized databases on the cheap. We’ll see how it affects their future.

Also, the Alabama .Net Code Camp this past weekend went great. I’ll have a post tomorrow summarizing things and posting a link to the files from my presentation.

Though it didn’t seem to get noticed much in the main technical press, the acquisition of Innobase by Oracle is big news in the open source database market. Many of you may recognize Innobase as the company that produced and supports InnoDB (one of the underlying database engines used by MySQL).

This could have dire consequences for the business model of MySQL AB. MySQL AB sells commercial licenses for their MySQL database product along with services and support. The contract regarding inclusion of InnoDB with MySQL is up for renewal next year and with Oracle holding the reigns now the negotiations could be difficult (and costly) for MySQL AB.

The press release by Oracle says they are committed to expanding the product and contributing to the open source community, but if you have seen what is to become of the PeopleSoft and JD Edwards products Oracle recently acquired, you may not be so convinced of their noble intentions.

For the open source crowd, we can take solace in the fact that InnoDB is GPL software, but with Oracle now “owning” the main developers on the product, new features and maintenance could dry up considerably. We’ll just have to wait and see what Oracle’s plans are for the newly acquired company.

In addition to MySQL, there are other alternatives to the high cost databases out there. Some of these are: PostgreSQL, Sybase ASE on Linux (Free for one CPU box, 5 GB max data storage, and 2 GB max RAM), Apache Derby, and FireBird

Here are some of the reactions and reporting from the Blogosphere:
KasLog – Oracle’s Innobase Acquisition; Open Source is Not Necessarily “Free”

Oracle’s Open Source

IT Toolbox – InnoDB and the Compromise of Dual Licensing

SAP Ventures – Oracle’s Open Source Database Endorsement

CIO Weblog – Oracle adds open source to the shopping cart

As Oracle guru Tom Kyte says of his daily Oracle experiences, I have learned something new about Oracle today. This trick works in Oracle databases 9i and up (like a fine wine this one will actually get better as you move to newer versions).

From a fantastic post on asktom.oracle.com entitled Can there be an infinite DUAL? I found a great alternative to using all_objects to generate X number of rows in a table.

Oftentimes, say for a Pivot Table, we will want to generate a table with X number of rows which we will Cartesian Product with some other table. To do this using all_objects (or some other table that is guaranteed to have a sufficiently large number of rows) it would look like this:

select rownum rn
from all_objects
where rownum < 10;

rn
–
1
2
3
4
5
6
7
8
9

9 rows selected

Now using a cool trick with 9i and up we can produce the same thing like so:

select l
from (select level l
from dual connect by level < 10
);

l
–
1
2
3
4
5
6
7
8
9

9 rows selected

Why do we care if the results are the same? The consitent gets for the second approach are much lower and they scale fantastically. Guru Kyte says that this will be even faster in 10g thanks to something called FAST DUAL. If you don’t believe me try it for 10,000 rows and check out the trace for both approaches.

Very cool stuff. Also did you know that Tom Kyte is blogging now at tkyte.blogspot.com? Very interesting to see a different side of this technology wizard.