Scenario – data transfer from system ‘B’ to system ‘A’ on request of system ‘A’
1. System A makes a remote function call to ‘B’ using synchronous RFC function
2. an asynchronous RFC function call from ‘B’ to ‘A’ using destination ‘BACK’ with a small size of data written back to ‘A’, continuing data collection in ‘B’
CALL FUNCTION ‘ZRFC_A_TO_B’
Pseudo code for ZRFC_A_TO_B
from table IM_ITAB.
1. BACK refers to the calling FM the same path as the called
2. The process need not be required to start from the beginning in case of communication failure during the runtime, as database commit occurs at each Remote function call
3. The processing time to write the data back to receiver system can be reduced
This blog attempts to take the idea presented in the blog on Parallel Processing – An Introduction one step further. Typically in case of any mass-processing activity (posting IDocs, for example), It is possible to leverage the processing power of multiple application servers. Since the processing of each record is usually independent of the processing of another, we can trigger multiple jobs on various application servers, each of the jobs processing only a part of the total records.