Hadoop gets native C/C++ injection
Big data for the non-Java generation
Like Hadoop but it’s just a bit too Javary? There’s now an answer for that: MapReduce for C (MR4C).
MR4C is an open source framework letting you run your native C and C++ code on the big-data crunching framework. Released quietly earlier this month and picked up here, MR4C lets you run C/C++ apps on Hadoop without needing to write your own special libraries.
MR4C was developed by satellite-picture specialist Skybox Imaging and has been flagged up by Google’s open-source team.
According to Skybox, it wanted to leverage Hadoop’s scalable data handling but access the existing ecosystem of image processing libraries developed in C and C++.
Despite their age, both retain their position as two of the world’s most popular programming languages.
Moreover, important here, C and C++ are popular choices in graphics-intensive, natively coded apps that talk straight to the underlying hardware.
Hadoop is the implementation of Google’s MapReduce, from Doug Cutting written in Java that’s seen as a platform for building businesses in addition to crunching vast amounts of data.
MR4C works by storing your C.C++ code as native stored objects with commands and inputs configured using JSON.