The DataSerializer implementation uses the {@link com.ibm.websphere.objectgrid.plugins.io.datadescriptor.DataDescriptorFactory} to create a DataDescriptor and populate it with attributes. After the DataSerializer is initialized, the eXtreme Scale runtime will retrieve the DataDescriptor from the DataSerializer to determine what attributes are available.
The MapSerializerPlugin implementation uses the {@link com.ibm.websphere.objectgrid.plugins.io.datadescriptor.DataDescriptorFactory} to expose a MapDataDescriptor, which identifies any relationships between maps.