db-derby-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Knut Anders Hatlen <Knut.Hat...@Sun.COM>
Subject Re: Australian JUG Presentation about Derby
Date Tue, 02 Sep 2008 12:13:42 GMT
Matt Doran <matt.doran@papercut.com> writes:

> Hi All,
>
> I recently did a presentation about Apache Derby at the Australian
> Java User's Group meeting in Melbourne.  The presentation was titled
> "Using Apache Derby in the real world".  Its focus was not so much on
> how to use Derby, but were it can be used and how it stands up in real
> world applications.  We use Derby in our print accounting application
> called PaperCut, and feel that it's one of the "gems" of the Java
> community.  Even thought we offer a choice of MS SQL Server, Oracle
> and Postgres within PaperCut, we find that 90% of our 10,000+ users
> stick with Derby - the default embedded option.  The presentation
> highlighted this as well as gave an overview of the idiosyncrasies and
> system performance tweaks we've made over the years.
>
> I thought I'd shared the presentation with the list as some of the
> content might be useful if others want to present Derby at their local
> JUG!

Thanks, Matt! The presentation was very interesting to read! I added a
link to it from the presentations section on
http://db.apache.org/derby/papers/.

> PDF:
>   http://www.papercut.com/anonftp/pub/open-source/apache-derby/ApacheDerbyAJUGPresentation.pdf

>
>
> HTML:
>   http://docs.google.com/Present?docid=dcz4c74g_82nr92vgnt&skipauth=true
>
> The presentation also proposes a few potential ideas for future
> features.  That is, moving more towards zero administration, and a
> better out-of-the-box experience for all application domains.

You may already be aware of this, but since the presentation mentions
updating the index statistics, I thought I'd just mention that a manual
way of updating the statistics (without dropping and recreating the
index) has been implemented and will be available in Derby 10.5, and
there is also work in progress to make it automatic. See
https://issues.apache.org/jira/browse/DERBY-3788.

> We think Derby's best "feature" is it's wide domain of applicability - 
> that is it can be used from anywhere from say a very small embedded
> database like say SQLite, all the way up to concurrent database
> applications matching MySQL, etc.  We've found that the default
> configuration options are however stacked in favor of the "small
> embedded" use case, and some manual performance tuning is required to
> get performance up on large datasets and more complex queries.  It
> would be great if the out-of-the-box domain of applicability was as
> wide as possible and one idea presented to achieve this is the concept
> of profiles.  For example, a profile parameter could be added to the
> current JDBC connection string and this could be used to reconfigure
> defaults more appropriate for the usage domain.

Sounds like a good idea. It would be great if you could log a JIRA issue
for this at https://issues.apache.org/jira/browse/DERBY.

-- 
Knut Anders

Mime
View raw message