Microsoft releases R Server 9.0 and free R Client 3.3.2

Microsoft today is announcing the release of R Server 9.0, an update for the commercially supported big data software aimed at large organizations. Microsoft is also releasing R Client 3.3.2, the latest version of the free software that people can run on desktop computers, as opposed to servers.

R Server was previously known as Revolution R Enterprise but was renamed earlier this year, following Microsoft’s 2015 acquisition of Revolution Analytics, a vendor of distributions of the R programming language.

[aditude-amp id="flyingcarpet" targeting='{"env":"staging","page_type":"article","post_id":2125727,"post_type":"story","post_chan":"none","tags":null,"ai":true,"category":"none","all_categories":"ai,big-data,business,dev,enterprise,","session":"A"}']

The new release — which draws on open-source R and community contributions as well as Microsoft’s own proprietary tools — contains a number of enhancements.

For one, the server software now includes the new MicrosoftML package with six new algorithms, including an artificial neural network for deep learning that can train on lots of data using graphics processing units (GPUs) and then make inferences on new data. (People can use one of many open source deep learning frameworks instead, but this is in R.)

There are also algorithms for more traditional forms of machine learning like random forest, boosted decision tree, and logistical regression, Microsoft partner director of program management Nagesh Pabbisetty wrote in a blog post.

With today’s release, models that data scientists write in R are easier to deploy into production. The server software lets data scientists expose models as application programming interfaces (APIs) and produce corresponding documentation using Swagger. R scripts and functions can be easily turned into web services as well.

R Client version 3.3.2 by definition is memory-constrained but can be connected to R Server. It contains some features that aren’t available in the community version, while at the same time supporting community packages. Currently it’s only available for Windows. It can run as part of Microsoft’s Visual Studio integrated development environment (IDE) with the R Tools add-in.

The new release can be installed offline, and once it’s installed, it checks for updates and tells users when they become available. The new MicrosoftML package is also included.