Can you decommission the InfoCube completely if you run BW 7.4 ?

[Former Member]

In classical BW environments, lots of InfoCubes are created with the sole purpose of performance, and are simply a 1:1 copy of the data in the DSO layer underneath it. So, Now in BW 7.4, by making a few minor adjustments to the DSO, and redirecting your Multi Provider (or Query) to the DSO instead of the InfoCube, can you decommission the InfoCube completely ?

[michael_devine · Accepted solution]

Hi,

Yes, absolutely. The read performance on a DSO and an in-memory optimised cube are comparable thus enabling the elimination of those 1:1 copies for read performance.

Please find the possibilities using LSA++

Regards,

Michael Devine

[Former Member · Accepted solution]

Hi Michael,

Thanks for the insight with the document you shared, much appreciated!

Regards,

Ashwin. 

[Former Member · Accepted solution]

Hi Michael,

We have evolved from SAP BI 7.0 are currently running BW 7.4. So, we are looking to

leverage the power of HANA DB and in this process we are trying to reduce the data footprint

in our current EDW layer to start with... and I'm trying to understand the options we have

to reduce the data footprint.

Regards,

Ashwin

[michael_devine · Accepted solution]

Hi,

If you really want to lower the data footprint check out Open ODS. Open ODS View - Using the SAP HANA Database - SAP Library

Regards,

Michael Devine

[Former Member · Accepted solution]

Hi Michael,

Could you please share your thoughts on the best way to handle year specific cubes in BW 7.4? We have around 4 Info cubes (year specific) based on a single DSO and we want to leverage new modeling techniques provided in BW 7.4

Current State :

1. Multi Provider <-- Info Cube 1 <-- DSO (main source) <--Extractor

                                  Info Cube 2

                                  Info Cube 3

                                  Info Cube 4                          

My Options are:

1. Multi Provider <-- Virtual Provider <-- Advanced DSO 1 <-- DSO (main source) <--Extractor

                                                              Advanced DSO 2

   2. Multi Provider <-- Virtual Provider <-- Cal View 1 <-- DSO (main source) <--Extractor

                                                             Cal View 2

                                                             Cal View 3

                                                             Cal View 4

Kind Regards,

Ashwin

[jdun · Accepted solution]

Why not just report directly off of the DSO?  Extractor->DSO->Query.  Multiprovider can go away since you no longer need to union the data (should be using composite providers anyhow).  Depending on your requirements calculation views may not be necessary unless you need to bring in non-BW data.  A lot of the complex calculations (including exception aggregation) are already processed at the HANA DB.