Data Mining Architecture
Applies to: 
SQL Server 2019 and earlier Analysis Services 
Azure Analysis Services
Fabric/Power BI Premium
Important
Data mining was deprecated in SQL Server 2017 Analysis Services and now discontinued in SQL Server 2022 Analysis Services. Documentation is not updated for deprecated and discontinued features. To learn more, see Analysis Services backward compatibility.
This section describes the architecture of data mining solutions that are hosted in an instance of SQL Server Analysis Services. The topics in this section describe the logical and physical architecture of an SQL Server Analysis Services instance that supports data mining, and also provide information about the clients, providers, and protocols that can be used to communicate with data mining servers, and to work with data mining objects either locally or remotely.
In general, SQL Server Data Mining operates as a service that is provided as part of an SQL Server Analysis Services instance running in multidimensional mode; therefore, we recommend that you also review the following sections of Books Online that describe the operation, maintenance, and configuration of SQL Server Analysis Services multidimensional solutions.
Processing a multidimensional model (Analysis Services)
Switch an Analysis Services database between ReadOnly and ReadWrite modes
For more information about how you can implement data mining in your business intelligence solution, see the Solution Guides section of the MSDN Library.
In This Section
Logical Architecture (Analysis Services - Data Mining)
Physical Architecture (Analysis Services - Data Mining)
Data Mining Services and Data Sources
Management of Data Mining Solutions and Objects
Security Overview (Data Mining)