Linux Databases Part II

Last week's Linux RDBMS overview left out several tools and many

readers quickly asked why certain RDBMS weren't listed. Space limit is

one reason and I also have to admit I'm not familiar with each and

every RDBMS around -- particularly those started on other platforms and

were ported to Linux only recently. This week I will try to make amends

and discuss four more RDBMS tools. Another issue, referring to my

statements that some RDBMS weren't suitable for large databases, has

been raised asking, "How large is large?" Lacking an industry

standard , common agreement dictates a large database contains more

than 4 GB of data or more than 4 GB of records -- 4 GB is the upper

limit of 32-bit integers and pointers. An RDBMS capable of storing

larger amounts of data are designed very differently from medium size

or desktop databases, and thus requires sophisticated storage and

lookup algorithms. Most Open Source products are still limited in this

regard. Now back to our overview.


Mini-SQL, or mSQL (not to be confused with MySQL), is a lightweight

RDBMS capable of providing rapid data access with little overhead. The

licensing policy is somewhat vague, but it's free for non-commercial


Pros: Compact, low resource usage

Cons: Unsuitable for large systems


Informix recently purchased the uniVerse RDBMS. uniVerse supports many

important standards and environments, e.g., ANSI SQL92, SQL3, ODBC,

Perl, etc.... Scalable, reliable, and easy to maintain,. uniVerse can

be accessed through various languages and tools.

Pros: Internationalized, resource efficient

Cons: Expensive, overkill for small databases


Borland released the source code and binaries of its vintage Interbase

RDBMS about a year ago. Interbase supports large databases (up to 32

Terabytes) and complies with the SQL92 and Unicode standards. It

offers "Event Alerters", a unique feature that similar to traditional

triggers, automatically notifies interested parties when specific

changes occur in the database.

Pros: Open Source, supports triggers, multidimensional arrays and


Cons: Linux support is still crude


Kdb offers outstanding processing speed (according to KX -- one million

updates per second) due to a unique architecture of "inverted" tables,

i.e. storing data together in each column, instead of the row

orientation used by other RDMS. Kdb is free for non-commercial users.

Pros: Extremely fast, very concise, in-memory database

Cons: Unique design that dictates oddities, e.g., proprietary


ITWorld DealPost: The best in tech deals and discounts.