When you use the new data flow, the PSA is used as the entry level for data in BW. During the loading process, the data is posted to PSA tables updated that have been created for activated Data Sources. The PSA is managed using the Data Source.
The previous PSA technology was based on the transfer structure. The PSA table is generated for an active transfer structure in this case. Prior to SAP Net Weaver 7.0, the PSA as a standalone application was managed in an object tree of the Administrator Workbench. The functions that you could previously execute in the PSA tree have now been replaced by functions in other trees (such as the Manage context menu option for a Data Source).
More information:Persistent Staging Area.
General Practices with Migration
When you switch to the new data flow and migrate the 3.x Data Source, an existing PSA for this 3.x Data Source is copied to the migrated Data Source with the previously loaded requests. XML and DU Connect 3.x Data Sources are exceptions to this. Since the switch involves a copy, a new PSA is generated for each Data Source upon activation.
After migration, you cannot use the Read PSA and Update Data Target process types any more. This must be replaced by the Data Transfer Process process type.