* It has a tiny fraction of the power of real projects like Hadoop, Storm, Spark, and so on.
* I created it because I wanted a simple way to run maps, reduces, and filters over many machines without making big changes to my original Clojure code. For this simple use case, many distributed computation systems seem quite complicated, so I thought it would be fun to write something extremely simple.
* I was inspired by how PigPen https://github.com/Netflix/PigPen offers a very natural interface to Clojure. I wondered if that could work on top of something simpler than Hadoop + Pig.
1 comment
[ 5.3 ms ] story [ 10.3 ms ] thread* I have not tested Kuzu in production yet.
* It has a tiny fraction of the power of real projects like Hadoop, Storm, Spark, and so on.
* I created it because I wanted a simple way to run maps, reduces, and filters over many machines without making big changes to my original Clojure code. For this simple use case, many distributed computation systems seem quite complicated, so I thought it would be fun to write something extremely simple.
* I was inspired by how PigPen https://github.com/Netflix/PigPen offers a very natural interface to Clojure. I wondered if that could work on top of something simpler than Hadoop + Pig.
* This isn't the first time something like this has been done. See net-eval for example http://nakkaya.com/2010/02/16/net-eval-dead-simple-distribut....
* Kuzu may or may not be useful to you, but I'd be interested in any and all commentary.