BW on HANA - Possible use with query Bex and/or BO 4.x

[Former Member]

Hi experts,

i wanted to ask a few things about BW systems on HANA for use query Bex and/or BO 4.x and use the in-memory for fast performances and not classic

System BW (poor performances):

1) The Bex queries and workbooks 7.x are supported on MULTIPROVIDER: I can directly use them without doing anything or do I have to do something to use in-memory on HANA ?
2) I can run directly Report BO 4.1 on BW on HANA with Query Bex on Multiprovider with Cube optimized for Hana or I have to create a snapshot query in the Hana Studio for the fast performances and use in-memory technology ?

3) What is the difference at performances query level if this query read the data from Infocube optimized HANA (BW Schema managed) and if the same schema of the cube I create a native HANA level as Calculation or Analytics View? Why should I create the schema in the native level if I already have the cube?

4) What is the necessity of export of model BW InfoCube into HANA design ?

Many thanks.

Charly

[Former Member]

Some of the questions are already discussed. Please check the link below

http://scn.sap.com/thread/3441016

[Former Member]

Hi Durgamadhab,

sorry, but i read the document "SAP BusinessObjects BI 4 x - Comparing interface options to SAP NetWeaver BW  SAP NetWeaver BW on SAP HANA" and don't understand why i may generate query snaphot for use BO 4.1 for BW on HANA and don'y can use BO 4.1 directly on Bex query 7.x. Can you answer this my question ? Is very imporatnt for me for possible update to BW on HANA. Thanks.

[Former Member]

Hi Charley,

There is three option to connect to BW on HANA..

1. You can just create a regular BEx and then consume it from BI 4.1 Webi using direct BICS connection using a OLAP source (this will be a little slow as there is BW layer in between)

2. You can expose the DSO of the BW as a BW Analytic view in HANA and then build all your logic in HANA studio and consume it in Webi. Creating all the logic like RFK and CFK will be a little trickythough through HANA unless you have a lot of HANA modeling experience. This will be the best performance but depending on your experience and requirement will need little more time for development

3. You can expose the BEx query firectly in HANA using RSDDB transaction as a HANA based query snap shot which preserves the RFK , CFK and it can leverage the BW multiproviders using this method.. This has minimal development effort if you have the BEx ready with good performance..There are very few limitation in this method however...

I hope this clears your doubt..

[Former Member]

Hi Durgamadhab,

many thanks for your ansers.

So, if I understand correctly, if in the document  "SAP BusinessObjects BI 4 x - Comparing interface options to SAP NetWeaver BW  SAP NetWeaver BW on SAP HANA"  in the column BW on Hana generated says to use a snapshot query is to have a good level of performance with respect to BW on RDBMS but is not required because I can always also use the 2 options that you explained for me. In practice, for the same amount of data and number of cubes on MultiProviders, made 100 made the index of ​​the response speed with a new model of native HANA (Analytical and Calculation View with CFK and RFK), the index will be 70 if I use a query snapshot and say 50 if I use the BICS connection between BW and BO?

In practice, if the quer Bex has also CFK and / or RFK to get the maximum performance from a system with cubes BW on HANA optimized for HANA, or use the snapshot query with a minimum of development effort or create a new schema natively on HANA Studio (Analytical Calculation and view) on which I create a BEx query or connect directly BO 4.1 on this view created in HANA. It 'just what I said?

Thanks. Bye. Charly

[Former Member]

The best performance will always be with the option 2 explained earlier but if you have a lot of Calculation and BEX variables as well as restricted Key figures you will need to explore the query snapshot option (option 3)..

Even in our BW on HANA implementation we use the option 2 to access the data from BW.

For query snapshot option the performance will be very good and very close to the option 2 so performance is not the only consideration for while choosing between query snapshot and direct access to the DSO rather on the complexity of calculation and existing BEx availability and so on..

[rama_shankar3]

With option#2: Having too many CKF and RKF in the BEx can have adverse effect in the BEx query performance. Remember, the OLAP processor has to run multiple SQL internally to restrict and calculate data. If you have too many CKF and RKF, you may want to model flags in the BW data model to improve performance and use flags in BEx directly.  With BW on HANA, using a DSO is sufficient but still modelling the logic is key to optimal use of HANA performance.

Remember, if you do not model right - even BW running on SAP HANA DB will result in poor performance. Please follow best practice for modelling complex data models.

I will shortly provide more best practise tips for BW on HANA modelling.

Regards,

Rama

[Former Member]

Hi Durgamadhab,

I'm sorry, one last thing I want to ask: let's say you choose option 2  (DSO import in Hana and creation of an analytical view and calculation view) or option 3 (Analytical View from snapshot query), the use these views are accessed to HANA directly from Query Be 4.x through 7.x by BO Web Intelligence, Xcelsius Dashboard and how (via JDBC / ODBC, etc.) ?

Thanks.

Regards.

Charly

[Former Member]

Hi Rama,

I'm sorry, what do you mean model flags in the BW data model to improve performance? bring the logiva of CKF and RKF directly in DSO?

Thanks.

Regards.

Charly