Outlining Layers in SAP Datasphere

When you want to set up a layered modeling approach, the following layers can be used:

1. Inbound layer: The foundational data ingestion layer that serves as the entry point for data from various source systems, providing direct 1:1 mapping from source systems with minimal transformation.

2. Harmonization layer: Objects on top of the inbound layer with a strong focus on re-usability, data standardization, semantic enrichment, and creating reusable views on top of the inbound layer with common business terminology and unified data structures.

   • In this layer, the individual data sets from different systems get harmonized in terms of standardized semantics and values.
   • The harmonization layer is often portrayed explicitly as the preparation layer for the objects in the propagation layer.
   • Use projections (column selection, column renaming) to harmonize field names from different systems and to apply static technical filters.
   • Use data access controls to restrict users from accessing all data.

3. Propagation layer: Objects on top of the inbound or harmonized layer, which provide semantic and value standardization of data from various sources, offering data in a highly harmonized form for further distribution and multiple use.

   • In this layer the individual data sets get combined with joins and unions, data is aggregated if required and dynamic filters are applied.
   • Master data objects and their semantics are modeled in this layer including text and hierarchy associations.
   • Data in this layer is consistent, complete and generally free of redundancy (except for performance reasons or complicated modeling tasks).
   • The propagation layer can also be used to serve the persistent architected data marts and is a basis for the reporting layer.
   • Use data access controls to restrict users from accessing all data.

4. Reporting layer: Typically objects (analytic models) that are accessed by reporting tools such as SAP Analytics Cloud.

   • Variables, exception aggregation, restricted and calculated measures are modeled here.
   • Use data access controls for further restrictions related to reporting on data sets.

For the sales scenario, three sources are used:

• SAP BW/4HANA: An SAP BW/4HANA system offers InfoObjects with master data for products and customers and a DataStore Object (advanced) with sales orders.
• CSV files: Two CSV files offer sales orders and product data.
• SAP S/4HANA: An SAP S/4HANA system offers a CDS view with sales orders.

The sales scenario is based on the Enterprise Procurement Model (EPM), which is an SAP template application in SAP S/4HANA. This scenario is based on a trading company called Itelo, a fictitious company that buys and sells computers and accessories. ITelO is a global player, with several subsidiaries and locations worldwide. It sells its hardware products through direct distribution channels. The company has various resellers and standard customers, as well as different suppliers.

Transaction code SEPM_DG has been used to generate mass data. This tool enables the simulation of real-world data volumes for purchase orders or sales orders and related master data, such as products, customers, suppliers or employees.

In the sales scenario, sales order extraction is based on DataSource ZEPM_SALES. This DataSource, which is not delta enabled, is based on a database view EPM_V_SALESORDER, which integrates the following EPM model tables from SAP S/4HANA:

• SNWD_SO: Sales order headers with 10,000 records for the years of 2020 and 2021.

• SNWD_SO_I: Sales order items with 58,800 records.

• SNWD_BPA: Master data for business partners with 46 records related to customers.

• SNWD_PD: Master data for products with 152 records.

For this scenario, CSV import is used for the CSV files and connections are defined for SAP BW/4HANA and SAP S/4HANA. Remote tables are used for data provisioning from SAP BW/4HANA, also including a remote table with real-time replication. For the loading of SAP S/4HANA data, a replication flow is used.

For harmonization, a view is defined on SAP BW/4HANA sales orders, a transformation flow is used for SAP S/4HANA data, and an intelligent lookup enriches sales order data with product data from CSV files. A task chain, including a data flow, is used to store this result in a local table.

In the propagation layer, harmonized sales orders from the three sources are combined and master data is associated.

For reporting, an analytic model is created, to which an data access control is assigned for authorization requirements.