OTMA Reroute wastes CPU

We have recently replaced the use of the LU62 Adapter with IMS Connect and we have noticed a large increase in CQS CPU consumption.

Our LU62 solution mostly involved the non-response mode protocol being used even though most of our applications actually reply via the IOPCB. Replies were queued to the LTERM and delivered asynchronously to the client system. In order to keep everything as much the same as before our IMS Connect solution also uses non-response mode. Most transactions are sent in using Send_Only with a Reroute_Tpipe name specified. We have clients waiting on Resume_Tpipe requests for the reroute tpipes. Note that we also use SuperMember support so the resume tpipes can be active on any IMS partner.

When an IOPCB reply is inserted it is first enqueued on the tpipe associated with the originating client. IMS then tries to deliver the message to IMS Connect but as there is no client waiting for a reply then it has to re-queue the message to the reroute tpipe.

All of the processing from the original IOPCB enque to the reroute enqueue is a waste and can be avoided - IMS has enough information at ahe time of the insert to know that the IOPCB reply would have to be rerouted.