Twitter and Cloudera are open-sourcing Parquet: columnar storage format for Hadoop. Take a look! parquet.github.com
OH: "There are always two solutions to every problem. One is deprecated, the other is not ready yet."
"Wow, you did 5 days of work in 1 day? You are, like, a 5x engineer!" "No, I am a 5x estimator."
So we had a JVM, and we put Lisp on it, and now we can call Lisp from a Java library imported by a Scala project. In a JRuby shell.
Omega: Google's rewritten cluster resource scheduler, arguably "next step" after Mesos and YARN: eurosys2013.tudos.org/wp-content/upl…