Microsoft R Server, or Revolution R Enterprise (RRE) as it is formally known, for now, is basically a platform for advanced analytics designed to tackle big data statistics, predictive modeling, and machine-learning capabilities.
The open source
R is limited in capacity as it is “single threaded and memory bound.” Microsoft R Server comes in handy if you need to scale and speed up your R applications by distributing massive data and computation on hundreds of nodes.
The architecture of Microsoft R Server includes the following underlying technologies which make it capable of handling a massive amount of data and advanced analytics: ScaleR, DistributedR, and ConnectR. Read more about its capabilities and architecture here.
The Microsoft R Server capabilities will also be available in SQL Server 2016 through R Services.
It is available in the following flavors (source):
- Hadoop—Scale your analysis transparently by distributing work across nodes without complex programming
- Teradata database—Run advanced analytics in-database for seamless data analysis
- SUSE Linux—Enable advanced analytics while making the most of your open-source investments
- Red Hat—Bring predictive and prescriptive analytics power to your Red Hat environments
- Windows—R Server for Windows will ship as R Services in SQL Server 2016 currently available as a Community Technology Preview (CTP)
The Microsoft R Server Developer Edition is available for download on MSDN. Here’s how to install it.
Microsoft R Server Installation Prerequisite
Download and install Microsoft R Open (MRO) which is an “enhanced distribution of R from Microsoft.” The
R engine version as of this writing is
3.2.2 released on January 1, 2016. Here’s the download site: https://mran.revolutionanalytics.com/download/mro-for-mrs/
Run the executable and follow the wizard:
You might have noticed the
Important! flag in the Microsoft R Open download page that says “If you choose not to write to the registry keys during the installation of MRO for RRE, then you may be missing some functionality.”
So make sure to tick the checkbox on the last page of the installation wizard:
Complete the installation. This R distribution from Microsoft comes with a GUI called, intuitively, Rgui. You might be more familiar with the popular RStudio. Other popular R gui’s are Deducer and R Commander.
Microsoft R Server / Revolution R Enterprise Installation
Execute the installation file
Revolution-R-Enterprise-8.0.0-Windows.exe. As of this writing, this version/build is the Revolution Enterprise 8.0.0 (build 310). The installer will check for prerequisites on your machine.
I’m installing this on a fresh install of
Windows Server 2012 R2 (VM on Hyper-V). You might have different items in the prerequisite list. Please take note, though, that the wizard cannot install the .
Net Framework 3.5 SP1. Add this feature through the Server Manager.
Once all the prerequisites are installed successfully, the Revolution R Enterprise 8.0.0 Setup Wizard starts. Click next:
Accept the license agreement. You will, then, be given the option to load the Recursive Partitioning and Regression Trees (
rpart) and Trellis Graphics for R (
lattice) packages by default. I’d click yes, and move on.
At the end of the installation wizard, click Finish to complete the installation:
Microsoft R Server GUI
Here’s the splash screen and the GUI: