none
Suggestion: Change R Server, R Client and R Open License Model to a per package license model RRS feed

  • Question

  • I would like to suggest a change in the Microsoft R Server, R Client and R Open Licence Model:

    In my opinion, it does not have too much sense to apply the same licensing model (per core) that has SQL Server to Microsoft R Sever, because of the fact that CRAN's R Distribution (From the R Core Team) is free of cost.

    Therefore, an alternative, and in my opinion, really profitable business model for Microsoft R, could be achieved by making the following changes:

    A) Offer Microsoft R Server free of cost to everybody. Remove R Client and R Open distributions.

    B) Integrate R Tools for Visual Studio in the VS Xpress Edition (+ Community Edition), and offer both GUIs available through this webpage. Or even make a unique GUI for R based on VS, as SQL Server has, and make it the default GUI when installing R Server (and not the old fashioned actual one).

    C) In Microsoft R Server, make "Mrsdeploy" the only proprietary  free  R package Microsoft offers, so that R Server is really a Sever.

    D) In MRAN (Microsoft R Package Store), make each one of the Microsoft proprietary  packages available at a certain cost. Therefore, customers could pay for each package they really need, or even the cores such package would use.

    By doing so, R Server would reach a more widespread Enterprise audience, and Microsoft "R Package Store" (MRAN) would be always there for improving default CRAN packages when company need them for high performance analytics.

    In fact, even in R Tools for VS an iOS like "App Store" could be created for MRAN, with a visual interface for Microsoft R packages, reviews, etc., where customers could buy all the machine learning and data science Microsoft packages they need.

    Thank you. Hope my suggestion is useful.


    Thursday, August 31, 2017 4:37 PM

All replies

  • In fact, if R Open really had a brand new GUI based on VS only for R, it could even be distributed with Microsoft Office Professional or University for example.

    Thursday, August 31, 2017 6:49 PM
  • It would be nice also to remove R Client maximum use of two threads so that ScaleR-powered functions work properly in local installations. 

    Tuesday, October 24, 2017 5:08 PM