ms.author: davidph

As part of the installation process for [!INCLUDE[rsql-productnamenew-md](../../includes/rsql-productnamenew-md.md)], a new Windows *user account pool* is created to support execution of tasks by the [!INCLUDE[rsql_launchpad](../../includes/rsql-launchpad-md.md)] service. The purpose of these worker accounts is to isolate concurrent execution of external scripts by different SQL users.

ms.author: davidph

Use dynamic management views (DMVs) to monitor the execution of external scripts (R and Python), resources used, diagnose problems, and tune performance in SQL Server Machine Learning Services.

ms.author: davidph

Data science and machine learning algorithms are computationally intensive. Depending on workload priorities, you might need to increase the resources available for data science, or decrease resourcing if R and Python script execution undermines the performance of other services running concurrently.

ms.author: davidph

SQL Server Trusted Launchpad service supports external script execution for R and Python. On SQL Server 2016 R Services, SP1 provides the service. SQL Server 2017 includes the Launchpad service as part of the initial installation.

SQL Server has an extensibility framework for running external script such as R or Python on the server. Script executes in a language runtime environment as an extension to the core database engine.

ms.author: davidph

The Python extension is part of the SQL Server Machine Learning Services add-on to the relational database engine. It adds a Python execution environment, Anaconda distribution with the Python 3.5 runtime and interpreter, standard libraries and tools, and the Microsoft product libraries for Python: [revoscalepy](../python/ref-py-revoscalepy.md) for analytics at scale and [microsoftml](../python/ref-py-microsoftml.md) for machine learning algorithms.

ms.author: davidph

The R extension is part of the SQL Server Machine Learning Services add-on to the relational database engine. It adds an R execution environment, base R distribution with standard libraries and tools, and the Microsoft R libraries: [RevoScaleR](../r/ref-r-revoscaler.md) for analytics at scale, [MicrosoftML](../r/ref-r-microsoftml.md) for machine learning algorithms, and other libraries for accessing data or R code in SQL Server.

This article describes the overall security architecture that is used to integrate the SQL Server database engine and related components with the extensibility framework. It examines the securables, services, process identity, and permissions. For more information about the key concepts and components of extensibility in SQL Server, see [Extensibility architecture in SQL Server Machine Learning Services](extensibility-framework.md)].

This article describes data collection methods you should use when attempting to resolve problems on your own or with the help of Microsoft customer support.