Hello BW enthusiasts!
In one of my earlier blogs, I put a spotlight on BW Transformations in HANA Runtime from ODP sources. The trigger came from lessons learned during an in-place conversion to SAP BW/4HANA were my team run into situations causing some headaches. The reason was, that there is no HANA Runtime on ODP sources in SAP BW 7.x. This missing HANA Runtime is available in SAP BW/4HANA again based on the assistance of the so-called Data Transfer Temporary Storage (DTIS).
However, the data is transferred to DTIS from ODP DataSources in serial mode only – which might lead in inferior total processing time compared to the classic ABAP Runtime which runs in parallel by default. As a result, some customers had to change back to ABAP Runtime or used PSA-like staging ADSOs to provision the data first into SAP BW/4HANA, before being able to put HANA logic on it.
The good news of today is, that this restriction has been removed:
One of the exciting new features of SAP BW/4HANA 2021 SP04 (released today / December 5, 2022) is the enhancement of the data extraction to DTIS: Now you can fill the DTIS in parallel from ODP DataSources. DTIS is now ready to receive large amounts of data in short time and DTPs can now make use of transformations in SAP HANA Runtime, and additional features like grouped extraction or error handling to update the incorrect records in the error stack. You can thus load the data into further data targets in a parallel and memory-efficient way, without the need for an ADSO as an intermediate layer, for example.
Preconditions:
The parallel filling of DTIS can be switched off via maintaining customizing view RSBKADMIN (tr. SM30). For example, the following entries prevent parallel filling of DTIS by a dedicated DTP:
– RSBKADMIN-FIELD = NO_PAR_DTIS_FILL
– RSBKADMIN-VALUE = DTP technical name
You can also switch the feature off for all DTPs based on:
– RSBKADMIN-FIELD = NO_PAR_DTIS_FILL
– RSBKADMIN-VALUE = ALL
Note: In scenarios where the extraction from the source system into the ODP Queue is the performance limiting factor, the parallelization of filling DTIS will hardly improve the overall performance. Performance benefits can only be expected in the processing of data from the ODP Queue into the SAP BW/4HANA InfoProviders based on the SAP HANA Runtime of the transformation logic.
Summary
The purpose of this blog post is to follow-up on my earlier blog from November 2021 providing the crucial update, that former restrictions have been removed with the latest SAP BW/4HANA Feature Pack (SP04). For more details refer to following sources:
– SAP note 3232927: ODP Data Sources – Parallelize filling of DTIS via Prefetch
– SAP note 3237249: ODP Data Sources – Parallelize filling of DTIS via Prefetch; PREREQUISITE
Appendix
By the way, this new SP of SAP BW/4HANA 2021 (release note 3220868) provides additional exciting enhancements, for example:
Okumaya devam et...
In one of my earlier blogs, I put a spotlight on BW Transformations in HANA Runtime from ODP sources. The trigger came from lessons learned during an in-place conversion to SAP BW/4HANA were my team run into situations causing some headaches. The reason was, that there is no HANA Runtime on ODP sources in SAP BW 7.x. This missing HANA Runtime is available in SAP BW/4HANA again based on the assistance of the so-called Data Transfer Temporary Storage (DTIS).
However, the data is transferred to DTIS from ODP DataSources in serial mode only – which might lead in inferior total processing time compared to the classic ABAP Runtime which runs in parallel by default. As a result, some customers had to change back to ABAP Runtime or used PSA-like staging ADSOs to provision the data first into SAP BW/4HANA, before being able to put HANA logic on it.
The good news of today is, that this restriction has been removed:
One of the exciting new features of SAP BW/4HANA 2021 SP04 (released today / December 5, 2022) is the enhancement of the data extraction to DTIS: Now you can fill the DTIS in parallel from ODP DataSources. DTIS is now ready to receive large amounts of data in short time and DTPs can now make use of transformations in SAP HANA Runtime, and additional features like grouped extraction or error handling to update the incorrect records in the error stack. You can thus load the data into further data targets in a parallel and memory-efficient way, without the need for an ADSO as an intermediate layer, for example.
Preconditions:
- The process is not controlled by a process chain in streaming mode.
- The ODP source system relies on an RFC connection.
- The DTP´s execution is done in parallel background processing (i.e. no simulation).
The parallel filling of DTIS can be switched off via maintaining customizing view RSBKADMIN (tr. SM30). For example, the following entries prevent parallel filling of DTIS by a dedicated DTP:
– RSBKADMIN-FIELD = NO_PAR_DTIS_FILL
– RSBKADMIN-VALUE = DTP technical name
You can also switch the feature off for all DTPs based on:
– RSBKADMIN-FIELD = NO_PAR_DTIS_FILL
– RSBKADMIN-VALUE = ALL
Note: In scenarios where the extraction from the source system into the ODP Queue is the performance limiting factor, the parallelization of filling DTIS will hardly improve the overall performance. Performance benefits can only be expected in the processing of data from the ODP Queue into the SAP BW/4HANA InfoProviders based on the SAP HANA Runtime of the transformation logic.
Summary
The purpose of this blog post is to follow-up on my earlier blog from November 2021 providing the crucial update, that former restrictions have been removed with the latest SAP BW/4HANA Feature Pack (SP04). For more details refer to following sources:
– SAP note 3232927: ODP Data Sources – Parallelize filling of DTIS via Prefetch
– SAP note 3237249: ODP Data Sources – Parallelize filling of DTIS via Prefetch; PREREQUISITE
Appendix
By the way, this new SP of SAP BW/4HANA 2021 (release note 3220868) provides additional exciting enhancements, for example:
- InfoObjects: New write-interface and OData interface, updated enhanced MD Update
- ADSOs: Better support for non-cumulative scenarios and new APIs for Semantic Groups
- DTO: NSE now available for Scale-Out-environments and much more
- DPP: Anonymization options
- BWMT: Enhancements for HCPR, Queries, DTPs, and a new Mass Activation View
Thanks to Karola, Horatiu and Claudia for your valuable input.
Okumaya devam et...