Hadoop was written in Java.  So it makes sense that Java will always be Hadoop's best friend.  Hadoop streaming is great tool that allows interoperability with other languages, but this still doesn't make other languages a 'first-class citizen' when it comes to working with Hadoop.

Microsoft thought it could unlock the big data world for the .NET developers with a project called Dryad, but no one ever used it.  So they put their backing behind Hadoop as well.  

I talked to Wayne Citrin, the CTO of JNBridge last week and he explained how you can't program directly in the APIs with Hadoop streaming.  Interop strategies using streaming also introduce additional overhead, and possibly errors, due to the necessity of parsing the stream .  

The engineers at JNBridge, however, are experts in .NET and Java interoperability (in either direction).  For a while they've been working on an idea using JNBridge technology for in-process interop.  

Today JNBridge has released Hadoop MapReducers in .NET as their new quarterly technology kit.  Key points below… 

( Download Here) ( PDF overview)














Check out the PDF overview of this technology.  ( Download Here)