Making R the Enterprise Standard for Cross-Platform Analytics, Both On-Premises and in the Cloud

  • Article

By Joseph Sirosh, Corporate Vice President, Microsoft Data Group

Less than a year ago we decided to acquire Revolution Analytics, the leading commercial provider of software and services for R, the world’s most widely used programming language for statistical computing and predictive analytics. At the time, we committed to building R and Revolution’s technology into our broader database, big data and business intelligence offerings and to bring these benefits to customers and students – on-premises, in the Azure cloud, and to new platforms.

We have since delivered many innovations and updates to help customers and partners benefit from the power of R:

Today, I have a few more exciting announcements to make about delivering R -based analytics to new platforms, developers and the R community:  

Delivering Microsoft R Server across multiple platforms allows our enterprise customers to standardize advanced analytics on one core tool, regardless of whether they are using Hadoop (Hortonworks, Cloudera and MapR), Linux (Red Hat and SUSE) or Teradata. For Windows, Microsoft R Server will be included in SQL Server 2016 as SQL Server R Services – and the combined bundle is less expensive than RRE standalone. Until SQL Server 2016 is released, Revolution R Enterprise for Windows remains available as a standalone product.

"Advanced and predictive analytics is about developing and testing new models. But it's also about their incorporation by developers into production deployments of decision support and automation solutions that can benefit the whole organization. With its new offerings for the R ecosystem, Microsoft is playing an important role in bringing analyst modeling and productivity tools as well as deployment tools to a broader audience," said Dan Vesset Program VP, Business Analytics and Information Management at IDC.

In addition to today’s announcements we'll continue to work on delivering greater integration and innovation in our upcoming offerings including:

These announcements reinforce our commitment to making it easy for enterprises, R developers and data scientists to cost-effectively build applications and advanced analytics solutions at scale, both on-premises and in the cloud.  

What is Microsoft R Server?

Microsoft R Server is a broadly deployable enterprise-class analytics platform based on R that is supported, scalable and secure. Supporting a variety of big data statistics, predictive modeling and machine learning capabilities, R Server supports the full range of analytics – exploration, analysis, visualization and modeling. By using and extending open source R, Microsoft R Server is fully compatible with R scripts, functions and CRAN packages, to analyze data at enterprise scale. It also addresses the in-memory limitations of open source R by adding parallel and chunked processing of data in Microsoft R Server, enabling users to run analytics on data much bigger than what fits in main memory.

Since the acquisition of Revolution Analytics, Microsoft invested in making the product even more safe, more international and accessible, and easier to install, while building a slew of new features (more details at Microsoft R Server – What is New?). For example, the current release includes features such as:

  • R Language version 3.2.2.

  • Enterprise class support that you would expect of Microsoft.

  • Enterprise level security conforming to Microsoft's Security Development Cycle, including threat modeling and attack surface analysis, code analysis, extended fuzz testing, and more.

  • Accessibility and compliance with Microsoft Accessibility Standards.

  • Support for the Chinese government standard GB18030 encoding.

With Microsoft R Server for Hadoop, customers who use Hadoop can build and run R models in a distributed cluster using Microsoft R Serverwhile working exclusively in their preferred R development environment. The software scales transparently by distributing work across Hadoop nodes, without the need for complex programming.

“At Hortonworks, we enable our customers to create modern data applications fueled by actionable intelligence from data in motion and data at rest. The combination of Microsoft’s R Server and our HDP and HDF platforms provides a scalable, enterprise-grade big data analytics solution based on the R language that millions of R developers can embrace for creating rich predictive analytics applications from billions of Internet of Anything data sources in a way that drives transformational value for the business," said Rob Bearden, CEO of Hortonworks.

With Microsoft R Server for Teradata, customers enjoy the advantage of bringing the analytics to the data by running advanced analytics models in-database on the Teradata appliances, rather than incurring the overhead of the traditional extract and analyze paradigm.

The Microsoft Data Science Virtual Machine will include a pre-installed and pre-configured version of Microsoft R Server Developer Edition, enabling R users to get started with data exploration and modeling right away on the cloud without needing to set up a fully configured system on premises.

Introducing Microsoft R Open

Revolution R Open is now called Microsoft R Open, and Microsoft continues its commitment of support for the open source R project, and to releasing regular updates to its enhanced, free distribution of R. Microsoft R Open enhances the performance of R with multi-threaded processor optimized computations provided by Intel Math Kernel Libraries (MKL) delivering large speedups especially in matrix oriented computations. It also makes it easier to build reliable applications with R on Windows, Mac and Linux by simplifying the management of R package versions. Microsoft R Open is 100% compatible with all R scripts and packages, and just like R is open source and free to download, use and share.

The R community is the keystone for the success of the R language, and a critical resource for data scientists, statisticians and now enterprises. Since the Revolution acquisition, Microsoft has continued to support the community, including expanding the sponsorship of R user groups and conferences. Investment in open source projects for R has also increased, with regular updates to Microsoft R Open and DeployR Open (a server for R deployment via web services), and new and updated R packages including checkpoint, an R package time-machine; ParallelR, for parallel R programming on clusters; and Azure ML, to deploy R functions to the Azure cloud as an API. In addition, Microsoft has pledged its support for the R Project by being one of the founding members of the R Consortium.

“As fellow members of the R Consortium, we’re thrilled to see Microsoft going all-in on support for the R language”, said JJ Allaire, CEO of RStudio Inc. “With a first-rate ecosystem of products and companies investing in R, the R language is truly enterprise-ready. We look forward to working with Microsoft on making customers successful with deployment ready, production quality tools and infrastructure.”

To learn more about Microsoft R Open, read the post by David Smith, R Community Lead. To get started and learn more about Microsoft R Server and Microsoft R Open, we invite you to sign up for our upcoming webinar series (dates below):

January 28, 2016

Intro to Microsoft R Open

David Smith

February 4, 2016

Using Microsoft R Server to Address Scalability Issues in R

Derek Norton

February 11, 2016

Data Mining with Microsoft R Server

Derek Norton

February 18, 2016

Best Practices for using Microsoft R Server with Hadoop

Jamie Olson

February 25, 2016

Using Microsoft R Server to Operationalize your Analytics

Jamie Olson

Here is a quick summary of all the links you need to get started:

As part of our commitment to help close the data scientist and analytics skills gap, we will also help educate and train aspiring developers and data scientists who want to learn R by leveraging the breadth and depth of our global programs and partner ecosystem, and with online courses such as Data Science and Machine Learning Essentials. I hope you will join us in our ambition to make R an enterprise standard for cross-platform advanced analytics, both on premises and in the cloud.

Joseph
Follow me on twitter