For some servers, it might be possible to provide one work area for XCF to use over and over for each new client request. This case might apply, for example, to a server whose request messages are of a known size and whose requests are processed synchronously by the server exit routine. If so, one might provide a work area when the server exit is called to initialize itself, and then never need to process a "get work area" request.

For some servers, a new work area must be provided for each client request to be processed. This case might apply, for example, to a server that arranges for asynchronous processing of the request by some other work unit. If the server exit leaves the work area available to XCF when it returns, the work area could be overlaid with the content of the next request message to be presented to the server exit, which in turn could corrupt the content of the work area as seen by the asynchronous work unit that is processing the previous request.

To preserve the integrity of the work area for the asynchronous work unit, the server exit needs to update the SXPL_WAD to either indicate that the work area is no longer available to XCF or to provide a new work area. Alternatively, the server exit could arrange for the asynchronous work unit to process a copy of the data in the work area, in which case the work area could be left intact for XCF to use with the next request.