LSA++ Simplified 

This is the modification that I am using in the design. We should avoid extracting the same data twice from the source to reduce load on source. Before ODP, since we have PSA, it make sense to extract data into PSA first and then do a parallel load to corporate memory and data propagation layer. If we are using ODP we can first load to staging to fetch data from source once and then to propagation layer. Another way to think if we want to fetch same data twice from the source, only pro I see is reducing batch time. We need to check both options based on our requirements.

@@balakrishnakatha8000 I've never seen a requirement for loading the same data twice. Even there's a case, that should be very very rare. I think SAP designed this concept as two options. Either from Source->CM->DWH or from Source->DWH. Not simultaneously Source->CM & Source->DWH.

@@ErdemPekcan The parallel load to staging ADSO is to store data in the source format 1:1 without any transformations. This staging ADSO is not used in daily flow but rather to support rebuilding scenarios meaning when there's an issue in the data consistency which requires us to load an entire year data or entire plant/material group data from source, now you don't need to put any burden on source but rather get the data from the staging layer ADSO. That's the main purpose of the staging ADSO. And yes, although data is stored twice but plz keep in mind that granularity is not usually the same meaning in staging ADSO, it'll be same grain as source but in Stnd or Datamart ADSO the grain is higher and matches the requirement. Additionally, please keep in mind that ADSO tables are column tables meaning huge compression takes place (~80%) so that too helps to keep memory in check.

@@DebanshuMukherjee sorry but my point and explanation was not about the purpose of CM and DWH layers. It was all about data loading concept.