Since living in Bosnia, there is often the need for displaying BH characters. This was the focus of the database comparison test. To get BH characters, there has to exist the support on the database level, connector (API), client-side controls, the Operating System, and the font used. Other interesting option is replication, which is useful for data backup and synchronization among several offices/departments.
|
color marks an unsatisfactory element, that eliminates the RDBMS as an optimal FOSS RDBMS. |
| RDBMS | Firebird 1.0 | Firebird 1.5 | MSDE 2000 | MSSQL Express | PostgreSQL | MySQL |
| URL | firebird.sourceforge.net | www.microsoft.com/sql/msde | www.postgresql.org | www.mysql.com | ||
| DL size | 3.94 (or 5.97 w/pdb) MB | 12.5 MB | 38.2 MB | |||
| Unicode support | Yes? Commercial connectivity needed for support. Workaround possible*. |
Yes | Yes | The documentation says yes but I
could not get Unicode characters in Navicat, for example. Connection character set was latin1. |
||
| Embedded server | Yes | No | ||||
| Replication |
IBRepl <-recommended Fibre FBReplicator |
Yes | subscriber only |
FAQ
states that only master/slave replication is available. Commercial options also available (not tested). Slony has limited options, no Windows instructions. |
Master-slave replication, other scenarios possible. Check here or here for a few advices. |
|
| API |
IBObjects, Zeos, UIB (Progdigy), dotNet data provider, ODBC |
.Net | any | ODBC dotNet |
dotNet ODBC Java |
|
| client-side controls |
IBObjects <- recommended, Zeos, UIB, FIBPlus, kbmMW, dotNet |
any | dotNet, Zeos, any, with ODBC |
dotNet, Zeos, any, with ODBC |
||
| license | free to use (IBPL) | free | free to use | free to use | free to use (BSD license) | free to use (dual license - GPL and commercial) |
| encryption | Yes, not tested yet | check FAQ | replication over SSL connections. |
Explanations:
Firebird: The documentation states that Unicode is supported but it is
hard to find API/libraries and client-side controls that support it as well.
Replication is available from the third parties.
Workaround for Unicode is possible by storing the data into tables with
character set None or Unicode_fss, or separate code pages. The font on the
client form is then used to display the appropriate characters.
MySQL: This RDBMS, or at least their web site, looks the most polished of the three. All tools are available on one place and all common options are available through these tools. MySQL also has the longest tradition of supporting both Unicode and replication so I assume that these are the best implemented.
PostgreSQL: It looks as if there is still no decent replication package available for PostgreSQL that works on Windows without a lot of trouble. The last thing a busy DBA needs, when a problem arises, is a software that makes more troubles. There are a few packages available for PostgreSQL. These are mostly old versions of commercial packages. Open Source versions are either in planning stage or are hard to setup on Windows.
MSDE 2000: This seems like the most complete solution among listed.
The conclusion is based on my current needs - replication and Unicode support.
It is also free for use. The limitations are max. database size of 2GB and
pipeline optimization for 5 concurent processes or 25 concurrent connections
(for web sites). I have not experienced this second limitation in number of
connections so can not give an opinion. The database size limit may eliminate
this option for larger databases.
http://www.microsoft.com/sql/msde/productinfo/features.asp contains detailed
explanation of all the features and limitations.
The upcoming SQL Server 2005 Express will support databases of up to 4GB, which
is more than enough for many organizations at the moment. The drawback is that
it will not be able to act as a Publisher and/or Distributor.
last updated 2010-08-10
Unique visits since 2006-10-15:
