In sap bw with hana, How exactly the performance is improved ?

[Former Member]

By replacing oracle db with hana , are we installing new processors as well. Or is the performance we get is purely because of columnar storage ?

[Former Member]

Hi Mohd,

This is a good question. Just replacing the DB or columnar store doesn't help.

If HANA DB is just a data base on memory then everyone can just add more memory for existing DB or columnar store then most DB already offer columnar store ( Like IBM DB2 Blu Accelerator or MS Sql columnar, etc) 

It is a combination factor of HANA DB which optimised for both read and write operations along columnar store and data compression. In case of HANA DB sap is trying to push most of the process to database.

All new BW modeling techniques ( ADSO or New Composite Provider) will work only with  HANA DB.

General points why BW on HANA is more efficient

1) No DIM ID in cube

2) Change log tables can be converted into views ( optional )

3) DSO Activation Pushed to DB

4) Transformation Pushed to DB if no ABAP routines.

5) No need of index for cubes

6) Aggregations is no more required.

Thanks,

Shakthi Raj Natarajan.