Explain the structure of data storage.
Speedminer does not store data in cube. This is because the database technology behind Speedminer - Caché is already an object and multi-dimensional architecture database, and thus we do not require to build cube (which from the name we can understand the purpose is to make 2-Dimensional RDBMS looks like multi-dimensional - Cube).

When ETL moves the data to OLAP database, even though from the table architecture and ODBC tools it seems like normal RDBMS database, but in fact, the data is already in multi-dimensional structure, and with our multi-dimensional indexing, bitmap indexing processing, index compression technology and our special processing for OLAP query, the statistics can be obtained in split second.

Since there are no need for building cube, the setup becomes extremely easy and user friendly, users will see OLAP database in the same architecture, same table relationship (in fact even field and every part is identical) as their operation database, and just need to select the fields required for analysis.

Is the tool able to integrate with Notes, SQL, Sybase at the same time? If yes, is there other tools that is required to make the integration work? If there is other tools, is it 3rd party or built in?
We have tested with SQL Server, Sybase and other ODBC compliance data source. Basically, any data source that allow access from outside their environment, we will be able to support, except minority cases like SAP which the customer has to purchase a separate ETL tools with special driver or purchase EAI (Enterprise Application Interface) from Hesper (Caché has an EAI product name Ensemble for such purpose). Generally, no other tool is required except in cases where special written driver is required like SAP.

更多理论请看 http://www.speedminer.com/faq.htm#about59