cancel
Showing results for 
Search instead for 
Did you mean: 

Avoid infocube n replace it by Hana optimized objects

0 Kudos

Hi Gurus,

We are using BW on HANA 7.4 SP7, currently in migration phase.

In the current flow we have (2)DSOs - >Cube->MP ->BEx queries(many).

As top down approach:-

Currently we have replaced MP with Composite provider and successfully copied the Bex Queries on top of new composite provider.

So as per standard practice we need to avoid the use of cube and use Hana optimized objects so any suggetsions? is it really standard practice?, if yes by which object we can replace cube? its SP7 so cant use advance DSO, can we use write optimized DSO? if yes by using it are we really achiving something (any improvement)?

Thanks in advance.

Snehal.

Accepted Solutions (0)

Answers (5)

Answers (5)

Former Member
0 Kudos

Hello Sneha,

As long as the union operations are performed, there is no runtime difference between a CompositeProvider and a MultiProvider, as the analytic manager is well suited for both.

You will have an improved performance if you're replacing infosets with compositeproviders.

Since you're in SP7 and you do not have access to ADSOs, I would recommend you to convert your standard infocubes to HANA optimized infocubes. This way, you would notice the improvements in performance.

Converting Standard InfoCubes to SAP HANA-Optimized InfoCubes - Modeling - SAP Library


Let me know for any clarifications.

Br, Harish

desgallagher
Contributor
0 Kudos

Hi Snehal,

It is not a requirement that you replace your info cubes with BW on HANA. The idea is that you replace them if you don't need them any more. Before HANA a lot of customers used info cubes as they could not report on DSO objects directly due to performance problems. So in the past a typical data flow might have been Data source >DSO >Info cube > Query , in this example dataflow the info cube only existed for performance reasons. With BW on HANA it is often the case the performance is ok when reporting on the DSO directly and therefore an info cube may not be required.

Kind Regards,

Des  

Former Member
0 Kudos

Hi Sneha,

In the Old BW, you were following LSA (Layered Scalable Architecture) set-up by having WO DSO's, Cube. But with New BW on HANA, you need to follow LSA++ set-up. With this, you can do away with Usage of Cubes & have Standard DSO's for your reporting needs.

Regards,

Swapnil

Former Member
0 Kudos

Hello Sneha,

If there is no Complex Transformations between 2 DSO to CUBE then you can use Virtual Info Provider than a Normal Cube.

At the same time check how much volume of data is extracting in the Bex query.

If the volume of extraction is high and all the Functions used in the Query is pushed to HANA then you can still use the Virtual Info provider else I suggest to go for HANA optimized info cube.

The condition which is you said in here has multiple alternatives and all are purely depend on the size of  Data , Transformations and Bex query functions.

Regards

Jagan

Former Member
0 Kudos

Hi Sneha,

Do you want to convert Standard Cube into HANA optimized cube? or you want to replace the hana optimized cube with other objects?

Regards,

Nikhil.