The purpose of this wiki is to help you increase your idoc inbound processing performance through the utilization of certain key reports and parameters.
The IDoc posting process is done in three steps:
1) Sending IDocs from partner system/ other SAP system to SAP
2) Receiving and storing IDocs in SAP data base
3) Processing IDocs in SAP system
Usually, this last step 3) meaning posting the IDoc data to the application is the one with the most workload. To increase performance here means increasing the overall performance. The entire process looks quite synchronous but isn't. And this is the key for performance tuning: Run this process in batch and in parallel.
This is how:
A) Switch from immediate IDoc processing to batch
B) Use parallel processing option
C) Use IDoc packets rather than single IDoc posts
Goto transaction WE20 in the SAP system (IDoc Partner Profile Settings) and make sure that the setting for the inbound partner profile of the affected IDoc message type is set to collect mode / background processing rather than processing immediately.
Now all inbound IDocs will stay in waiting position which means yellow with an IDoc status of 64 ("ready to be processed"). To process such IDocs you can use ABAP program RBDAPP01. This IDoc execution program is then to post the collected IDocs with a valuable package size and server group in parallel mode.
Goto transaction SE38 to start the ABAP program "RBDAPP01". You can run this in online mode or in background (as a scheduled job which is recommended). The key here is to use the right selection criteria:
- Use parallel mode (which can be switched on at the second tab of the selection screen).
- Use just the IDoc type and IDoc subset you need to process.
- Use a proper configured server group for the parallel load (done in transaction RZ12)
Finally, there is one more important value that needs be used as a selection criterion for the RBDAPP01 report. This is the so called packet size. See the following link for more information on choosing a package size that suits your needs:
Choosing the correct package size for RBDAPP01
Using these performance tuning tips, you should be able to increase the processing time and finally to load mass data. Just have a look the system performance during the runtime (transaction SM50/SM51, transaction SM58 and (if using the batch scheduler) SM37) where you easily can see whether there are any bottle necks.