Thursday, August 28, 2008

MapReduce Integrated into Relational Database

There are two different approaches when it comes to processing of very large data set in a parallel and distributed fashion. One - the traditional - preferes the relational database while the other - strengthen by the Web 2.0 hype - prefers more constraint and data structure relaxed environment represented for example by MapReduce approach.

Apache's Hadoop projects is the only available production ready implementation of MapRecudce and it has captured a lot of attention and its popularity is growing fast. No wonder that these two worlds run into flamed and little productive discussions concerning "which one is better?" question.

Well, the interesting news is that two relational database engines are now providing integrated solution of both the worlds:
News found via http://www.dbms2.com/2008/08/25/mapreduce-sound-bites/

I am particularly curious how soon will be the Hadoop integrated into Oracle or MySQL database.

No comments: