PDA

View Full Version : Making the case for Kettle



MattCasters
09-04-2007, 06:31 PM
Dear data integration fans,
Once in a while, there are discussions on various blogs (usually with me smack in the middle of it) debating the differences between code generation and model based execution, how this impacts the way we approach databases, the open nature of it all, etc.
With this blog entry I want to push the notion that Pentaho Data Integration (Kettle) (http://kettle.pentaho.org) didn’t just evolve by chance into the state it is today as a streaming, metadata driven, model based engine. I made some careful design choices early on…
Open as possible
The goal of Kettle from the beginning was to be as open as possible. My definition of “as open as possible” included:
<ul> open source with an LGPL license (see this JBoss link [PDF] (http://www.jboss.com/pdf/Why_We_Use_the_LGPL.pdf) for a nice explanation)
open, readable XML format
open, readable relational repository format
open API
easy to set up (