XI message synchronization problems

Question

We are having a problem with XI that is related to timing of transmissions from a Legacy Mainframe system to our R/3 system. This is the scenario:
Update 1 is transmitted from the mainframe at 1:00pm and contains an update for item “Y”. The update fails on the R/3 system due to a lock error so the XI system waits before trying to do the update again. Update 2 is transmitted at 1:05pm and contains an update which is also for item “Y” and the update is successful on the R/3 system. At 1:07pm Update #1 (that originally failed due to a lock on item “Y”) retries and it is successful. The result is that we have updates to item “Y” in R/3 that are out of sequence and in some instances puts item “Y” in an invalid state.
     I know we could resolve the problem by creating a serial queue but this is not an option because we have many updates for many items being transmitted via these queues. I need to know if there is a way for XI to know that item “Y” contains an update that was locked out (and is in a retry state) and to hold any additional updates for item “Y” until the first update is successful.

Answer 1

Hi Ed,
you could create a business process, which is receiving from Legacy Mainframe as long as no positive Acknowledgement came from last send message. That means you design a process where you receive (while 1 = 1) and send messages. If you receive a positive acknowledgement AND no other message is to be send, than stop the process.
Of course a BP does not upgrade performance at all
BTW: You should consider using EOIO...
Regards, Udo