At present 2 tier architecture is implemented in our dashboard .
we are trying to make it as 3 tier like
1)Extract(Incremental load )
2)DataModel( Load qvds and then do transformations (like joins,applymap,concatenations) based on Requirement )
3)Binary load .
Currently the dashboard size is 8 GB.
currently the qvds that are required for this dashboard are around 40 qvds.
All the qvds size is 22 GB.
I thought of 4 tier approach where in we will do all the transformation logic in 2nd layer using the base qvds(Extract layer) and in DataModel layer , load only transformed QVDS which will be of OPTIMIZED LOADS.
But this approach will make the duplication of qvds or Creation of additional qvds which may go around 15 GB.
