Show TOC

Function documentationParallelized Initial Data Transfer

 

In most cases, the initial data transfer runs in serial. For a serialized initial data transfer, the system selects all objects sequentially in the SAP ERP system, transfers them to the target system, and then processes them there in the same sequence. The correct sequence of individual objects is guaranteed because existing dependencies between the objects are taken into account. However, it may be useful to change to parallelized data transfer for performance reasons. You can activate several integration models in parallel; that is, at the same time. In the same way as for a serialized data transfer, the Core Interface (CIF) ensures that an object is only transferred once, even if it is contained in several integration models.

For certain object types within an integration model (see below), you can specify that the transfer of objects of this type is to run in parallel. For this, you can decide whether the selection in the ERP system, the subsequent processing in the target system, or both, are to run in parallel. However, parallelization is only possible within each object type; that is, the objects of one object type are transferred in parallel, but the object types themselves are transferred in serial as before. The system only begins to process the next object type once the entire selection for the previous object type has been completed.

Prerequisites

  • You have defined a number range in Customizing for CIF in the SAP ERP system for parallelized initial data transfer (see CIF Customizing in the ERP System). For more information, see Customizing for SAP ERP under Start of the navigation path Basic Settings for the Data Transfer Next navigation step Initial Data Transfer Next navigation step Determine Number Ranges for Parallelization. End of the navigation path.

  • The above setting is dependent upon the current target system. Since in most cases the production system has a different target system from the system in which Customizing is carried out, you have also made this setting on the SAP Easy Access screen of the productive system under Start of the navigation path Logistics Next navigation step Central Functions Next navigation step Supply Chain Planning Interface Next navigation step Core Interface Advanced Planner and Optimizer Next navigation step Environment Next navigation step Current Settings Next navigation step Define Number Ranges for Parallelization End of the navigation path.

Features

The parallelized initial data transfer incorporates the following functions:

  • Parallelized selection in the ERP system

    The ERP system selects the objects to be transferred in parallel in different work processes. For this, parallelization takes place with reference to the blocks of a filter object type. Each work process processes one filter object block. You define the block size (that is, the maximum number of filter objects processed in each block). For more information, see Customizing for SAP ERP under Start of the navigation path Integration with Other mySAP.com Components Next navigation step Advanced Planning and Optimization Next navigation step Basic Settings for Data Transfer Next navigation step Initial Data Transfer Next navigation step Set Filter and Selection Block Size End of the navigation path.

    Note Note

    The performance of the initial data transfer is determined not only by parallelization, but also by the size of the filter object blocks. If you choose a large block size, it may be that the system forms only a small number of blocks within a filter object type. This means that only a small number of work processes are parallelized. In contrast, a higher degree of parallelization is possible for smaller block sizes, although only a small number of objects are transferred in each block. If you want to optimize the performance of the initial data transfer, you have to take into account both parallelization and the filter block size and adjust these according to the specific requirements of your system.

    End of the note.
  • Parallelized processing of multiple queues in the target system

    The target system processes several incoming queues in parallel in different work processes. In this case, the system uses a different queue name for each data block that is sent using qRFC. The queue name contains the number of this data block.

    Parallelized processing can result in a significant improvement in performance. As before, you use transaction CFC3 to specify the block size (that is, the maximum number of filter objects processed in each block).

Parallelized transfer is currently supported for the following object types:

  • Material masters (in SAP APO only)

  • Production process models (in SAP APO only)

  • Sales orders (in SAP APO only)

  • All types of stock (in SAP APO only)

  • Planned orders and manufacturing orders (in SAP APO only)

  • Networks (in SAP APO only)

  • Maintenance orders (in SAP APO only)

  • Purchase orders and purchase requisitions (in SAP APO only)

  • Sources of supply (contracts, purchasing info records, scheduling agreements)

  • Batches (in SAP APO only)

  • Inspection lots (in SAP APO only)

  • Product data structure (PDS) (in SAP APO only)

    To use parallelized data transfer for PDS, you have to activate the Enhanced Initial Data Transfer (CA_CIF_INILOAD_01) business function, which is available as of SAP ECC Enhancement Package 4.

Note Note

For a parallelized initial data transfer, the sequence in which the integration models are activated is specified not by CIF, but exclusively by the time at which they are activated. For example, if a parallelized initial transfer of material masters and purchase requisitions is triggered, it may be the case that the material master data that is required for the transfer of purchase requisitions does not yet exist in the target system. It is therefore recommended that you only transfer transaction data once all the required master data has been successfully transferred.

Parallel postings to the same material/plant combination or resource can lead to a liveCache lock in the target system. In particular, this can be the case if two integration models for different object types (purchase requisitions and sales orders, for example) are activated in parallel and are both related to the same material/plant combination. This also applies to object types that contain subobjects, such as planned orders or manufacturing orders. As a result, you should ensure that integration models that are activated in parallel do not access the same material/plant combination at either header or subobject level or with reference to the resource used. Object types to bear in mind here are planned orders, manufacturing orders, networks, and maintenance orders with the operation and component subobjects, subcontract orders and subcontract purchase requisitions, and stock transport orders and stock transport requisitions.

End of the note.

Activities

As standard, the initial data transfer is strictly serialized. To activate parallelized initial data transfer, do the following.

  1. On the SAP Easy Access screen, choose Start of the navigation path Logistics Next navigation step Central Functions Next navigation step Supply Chain Planning Interface Next navigation step Core Interface Advanced Planner and Optimizer Next navigation step Integration Model Next navigation step Activate End of the navigation path or use transaction CFM2.

  2. On the Activate or Deactivate Integration Model screen, select the Parallelize Selection in ERP System checkbox.

  3. Select the Parallelize Processing in APO checkbox.