To support toolkit applications, a process must be able to access request data and output data into a response message. To do the inputting and outputting, the Abstract Layer uses request and response mapper to input data into contexts and output data from contexts. Note that the contexts can be local contexts or remote contexts (see CHA) depending on their definition in the BPEL file.
The Abstract Layer uses the appropriate object mappers for the object's type and attributes to map the formatted objects in the request message into the context. If the Abstract Layer does not find the appropriate mapper, it uses the Java(TM) bean mappers as the default.
At the end of a process flow, the Abstract Layer maps the context data into the response message. Because the response might not need all of the data contained in the context, the Abstract Layer uses the MapList custom property in the BPEL file of the process to determine which context data it will include in the response. Because the process can only use one part of the output message, the output part is a Map that contains the output data.