Apache Hadoop Get Together Berlin

2009-09-29 16:38
The Get Together started just a few minutes ago. The room is packed with more than 35 people this time. This is the first Hadoop Get Together in Berlin that will be recorded on video, thanks to Martin from newthinking for doing the recording and post processing as well as to Cloudera for sponsoring the videos.

The first talk was given by Thorsten Schuett on solving puzzles with map reduce. His disclaimer: Working at ZIB Berlin he had a large cluster in the basement to put to good use. However the cluster does not run Hadoop. It is based on Lustre FS and does not rely on commodity hardware. So he implemented a solver for 4x4 sliding puzzles in a map reduce framework targeted for "his" cluster.

Second talk was by Thilo Goetz on JAQL, a language for querying JSON documents that can run queries on top of a Hadoop cluster.

In the third and last talk, Uwe Schindler gave an overview of the new features and performance improvements of last weeks Lucene 2.9 release.

After raffling the Hadoop books donated by O'Reilly, we will move to a bar close by after the talks are over to have some beer and continue discussions. A summary that includes more details as well as links to the slides will be online soon.

Update: I had reserved a table at Cafe Aufsturz close to newthinking store for about 15 people - maybe less, maybe more. We ended up going there with more than 25 people - really glad there were still enough tables left for us :)

Update 2: Next meetup - December 16th, I already got one definite and two tentative proposals for talks.