Introducing Analysis Services Architecture - Data Mining

Analysis Services works in a simple client-server architecture, allowing clients to connect through a TCP/IP connection or using IIS to connect through HTTP through the Internet. Analysis Services provides a variety of clients, such as OLE DB, ADOMD.Net, and more. Rather than introducing a new communication mechanism for each client, each client is merely a thin shim that communicates with the server by producing and consuming the core Analysis Services interface, which is XML for Analysis (XMLA). XMLA is a simple XML protocol for communicating with analysis servers regardless of the source or destination platform. XMLA is covered in more detail in the next section.

When a request is received by the server, Analysis Services determines whether the request is an OLAP request or a data mining request and routes the request appropriately. In the end, a data mining request invokes a data mining algorithm that can be one created by Microsoft and included with the product, or can be a custom third-party plug-in algorithm loaded into the Analysis Services server. Additionally, a user-defined stored procedure can be invoked that uses a server-side version of the client ADOMD.Net interface to access server objects and model content directly.

A diagram of the data mining client architecture is displayed in Figure. The shaded boxes represent components that are part of Analysis Services, and, where appropriate, identify the files necessary to use the component.


All rights reserved © 2018 Wisdom IT Services India Pvt. Ltd DMCA.com Protection Status

Data Mining Topics