Show TOC

Background documentationActivation for MPP Architectures Locate this document in the navigation structure

 

For MPP databases, the system uses performance-optimized data in the DataStore object. This means the activation runtime can be significantly reduced for standard DataStore objects.

Note Note

If you upgrade from SAP NetWeaver BI 7.0 to Business Warehouse 7.3 (and if the database is MPP-capable), the existing DataStore objects are still processed using the standard activation. MPP processing is only performed after the reactivation of the DataStore object metadata. In the production system, you can only use report RSDG_ODSO_ACTIVATE to reactivate all the DataStore objects in the system if you write the changes to a transport request.

End of the note.
Supported Databases

The activation for MPP architectures is supported for IBM DB2 Universal Database for UNIX and Windows.

IBM DB2 for Linux, UNIX and Windows has to use the Database Partitioning Feature (DPF) and the DataStore object has to be distributed across at least four database partitions.

For more information on DPF, see the Database Administration Guide "SAP NetWeaver Business Warehouse 7.0 and Higher — Administration Tasks: IBM DB2 for Linux, UNIX, and Windows" in SAP Service Marketplace, under:  http://service.sap.com   instguidesnw   Your SAP NetWeaver Release   Operations   Database-Specific Guides.  .

Constraints

In the following cases, it is not possible to use activation for MPP architectures. The standard activation is used instead:

  • If activation is canceled where a standard activation has been used, the system performs a standard activation again.

  • If the Unique Data Records flag has been set in DataStore object maintenance.

  • If the number of InfoObjects exceeds a specific limit.

  • If the Do not condense requests into one request flag has been set in DataStore object administration or in the corresponding process chain variant in the Activation dialog box.

  • If the activation of unique data records only has been forced in the DataStore administration or in the corresponding process chain variant of the Activate dialog box.

  • If multiple requests have been activated together, the aggregation behavior of the InfoObjects must be identical.

Technical Process Flow of the Activation

Activation starts with the main process. A prepare process splits the data into data that is suitable for MPP architecture activation and data that is not suitable and therefore has to be processed using standard activation. Records that occur more than once (due to the semantic key) in the requests to be activated are processed using standard activation. It can therefore be helpful to activate multiple loaded requests separately, so that you can make better use of the benefits of MPP architecture activation. Records with record mode "ADD" are also processed using standard activation.

The prepare process fills a table where the records from the active table and the activation queue are added by Outer Join. Standard activation then starts on this Outer Join table. The runtime also improves because the data does not first have to be read from the active table. Activation for MPP architectures then starts and fills the table for active data and the change log.

Note Note

Activation for MPP architectures processes uses serial processing to process the new records, the records to be updated and the records to be deleted in sequence. The table of active data and the change log are updated in a database transaction. Depending on the amount of data and the data distribution method, this can place high demands on the rollback segments or the memory area for the active log files of the database. The system should be configured accordingly.

End of the note.