The import for IMS/TM takes a COBOL copybook and generates an XML adapter definition which is used by the AIS server in translating an input XML to a buffer that will be sent to the transaction. The flow of an incoming request is:

incoming xml -> AIS Server (translated to binary buffer) -> OTMA C/I -> IMS/TM

The return path of the response is identical. The "problem" with the above is that the COBOL copybook deals with what the transaction actually gets and replies. This is not exactly what the AIS server needs because the OTMA layer adds some prefix to the input and removes a prefix from the output. The differences are:

1. The LL-ZZ pair of fields are added by OTMA and should not be part of the AIS adapter definition.
2. The transaction name (and the space following it) are added by OTMA and should not be part of the AIS adapter definition.
3. On the output side, the output LL-ZZ pair are removed by OTMA and are not provided to the AIS server. These fields should therefore not be part of the output record.

The AIS import for IMS/TM will be modified to deal with these prefixes, but currently these extra fields need to be manually removed following the import process.