This section provides an overview of the architecture of SQL Server 2016 R Services, and of SQL Server 2017 Machine Learning Services.
The architecture for the extensibility architecture is the same or very similar for the SQL Server 2016 and SQL Server 2017 releases, and similar also for R and Python. However, to simplify the discussion, this topic discusses only the R components, including new components added in the SQL Server database engine to support external script execution, security, R libraries, and interoperability with open source R.
Additional details are provided in the links for each section.
Both SQL Server 2016 R Services and SQL Server 2017 Machine Learning Services (In-Database) install an open source distribution of R, as well as packages provided by Microsoft that support distributed and/or parallel processing.
The architecture is designed such that external scripts using R run in a separate process from SQL Server. Current users of R should be able to port their R code and execute it in T-SQL with relatively minor modifications.
To scale your solution or use parallel processing, we recommend that you use the RevoScaleR package or the MicrosoftML package. If you do not use the distributed computing capabilities provided by these libraries, you can still get some performance improvements by running your R code in the context of SQL Server.
For more information about the external scripting components that are installed, or the interaction of SQL Server with R, see R Interoperability
Components to support R integration
The extensibility framework introduced in SQL Server 2016 is continued in SQL Server 2017. The extensibility components are used by SQL Server to start the external runtime for R, to pass data between R and the database engine, and to coordinate parallel tasks needed for an R job.
The role of these additional components is to improve data exchange speed and compression, while providing a secure, high-performance platform for running external scripts.
For detailed description of the components that support R, such as the SQL Server Trusted Launchpad and RLauncher, see New Components.
When you run R code using Machine Learning Services or SQL Server R Services, all R scripts are executed outside the SQL Server process, to provide security and greater manageability. This isolation of processes holds true regardless of whether you run the R script as part of a stored procedure, or connect to the SQL Server computer job from a remote computer and start a job that uses the server as the compute context.
SQL Server intercepts all job requests, secures the task and its data using Windows job objects, and maintains security over data using SQL Server user accounts and database roles.
Data is kept within the compliance boundary by enforcing SQL Server security at the table, database, and instance level. The database administrator can control who has the ability to run R jobs, and who has the ability to install or share R packages. The administrator can also monitor the use of R scripts by either remote or local users, and monitor and manage the resources consumed.