IBM Db2 - The Ultimate Database for Cloud, Analytics & Mobile
DB2 10.5 has been made available recently. One of the most interesting features is Columnar based tables and the BLU technology. This article will help you to get up and running in no time with your first BLU table.
This week IBM made available DB2 10.5 for download and you can get a trial version of the product free of charge.
I have been looking after BLU tables and this was a great temptation to give it a try.
Indeed, one of the most relevant features of DB2 10.5 is the introduction of BLU Acceleration and Columnar tables. This is a new technology (to DB2 LUW) that has the potential to dramatically accelerate analytics queries. A key advantage is that BLU is embedded directly into the DB2 kernel and its implementations should be transparent for us, the users, apart of its performance advantages, of course.
Now we can create column organized tables, in contra-position to traditional row organized tables, in DB2 databases. DB2 10.5 embeds a vector processing engine in its core that takes in charge its handling. Some publications describe queries qualifying and touching 10 Terabytes of data being resolved in seconds, or less. A statement that is strong enough to keep many of us indifferent, right?
Another very interesting characteristic of Columnar based tables is simplicity. Whereas a row organized table requires the usual set of interactions that goes from designing the physical model to creating the best indexes for performance, BLU tables just need to be created and loaded. And that’s it. No more questions asked.
These notes are a recollection of the steps that I followed, and the problems that I encountered, while creating my first set of BLU tables. Hopefully these notes will help you to get started quickly with BLU as well.
Add a Comment