Standalone HANA Vs BW on HANA

[KulDeepJain]

Hi HANA Experts,

I am newbie working on my first HANA POC. 

We are having following system in client landscape:

1. SAP ERP
2. SAS
3. SQL Server
4. BODS as ETL
5. XYZ applications
6. No SLT licencing

We want analytics to do on HANA, data will be coming from all this servers

What will be best way to get data in SAP HANA system:

1. BW on HANA

2. Standalone HANA

What are the pros and cons for implementing of above architecture?

want a good disucussion here to understand what will be good in what aspects...

Message was edited by: Tom Flanagan

[former_member182302 · Accepted solution]

Firstly congrats on getting your First HANA Project.

And yes you given details on landscape, but using BW or not using BW depends on the functional requirements.

So you may want to be specific in your question.

Have a look on this blog from John on when we need BW or when Standalone HANA is enough

Blog: Does SAP HANA Replace BW? (Hint: No.) - P... | SAP HANA

Regards,

Krishna Tangudu

[KulDeepJain · Accepted solution]

Hi Krishna,

Thanks for your reply,

Its true we need to understand functional requirement to make decision.

I am keeping my point of view below,,,,

please look into this and provide your imputes.

Sr. No.	SAP BW on HANA architecture		Stand alone HANA architecture
1	Business content for faster and Consistent development		RDS solution for faster development
2	Self ETL with SAP and Non SAP sources (No separate licensing )		No itself ETL
3	Supports BODS for scheduled Batch data load		Support BODS with Near real time data load
4	Support Mixed scenario means BW modeling can also be Accessed in HANA studio		It also support Mixed scenario
5	Time dependent master data can Handle easily		Hard to handle Time dependent Master data scenario
6	Dynamic hierarchy to handle easily		Hard to handle dynamic hierarchy
7	Inventory scenario can handle easily		Can't Handle as of now
8	Virtual Info-cube means without Storing data reporting is possible		Not possible as of my knowledge
9	Easy to built complex modeling		SQL script knowledge required to achieve
10	Not Possible as of my knowledge		Real time reporting Via SLT
11	Standard BW authorizations		HANA based authorizations

I want to have healthy discussion here, please provide your views on above points and also add additional points to give more pros and cons here,

[Former Member · Accepted solution]

Hi,

I would not consider points 2, 3, 5 & 9 as much differentiation.

From tools perspective, if you buy enterprise HANA license you will get both SLT and BODS included in the bundle. Hence, the comparison becomes whether you want to use BW as your data warehouse application or you want to develop a custom data warehouse on HANA using BODS, SLT and HANA modeler. Here is my opinion based on my personal preferences and biases.

Advantages of BW as data warehouse:

1. Standard BW extractors are available for easier access to data that does not need in-depth understanding of SAP ERP entity relationships.

2. Modelling tools are easier to use

3. Data warehouse management is much easier (central monitoring and management of ETL process)

4. Easy to implement hierarchies.

5. Planning functionality (integrated planning) is available.

6. Full functionality of SAP Netweaver platform is available for custom development.

Disadvantages of BW:

1. BW does not conform to a puristic vision of a data warehouse in some practioners' mind. Please note that this will become a religious topic once you enter into this argument.

2. BW extractors can sometimes be a pain when you need to bring in additional fields and recapture historical data for the additional fields.

3. Reuse of KPIs can be problematic if they are defined using BEx query designer at a cube or multiprovider  level.

4. Needs BODS to get data from Non SAP sources (unless you want to use flat file upload)

Advantages of HANA (Enterprise license)

1. Complete freedom to develop data warehouse as per your wish (religion).

2. Powerful modelling tools

3. BW extractors in SAP ERP can be used  (DXC method) to get SAP ERP data.

4. Any KPI defined at analytic view or calculation view is visible in higher level calculation views. Hence, high degree of KPI reuse. This avoids duplication of KPIs .

5. Can easily integrate with any applications developed in HANA XS, Java, .Net or ABAP technologies using ODBC, JDBC, ADBC connectors.

Disadvantages of HANA

1. Expensive to license.

2. Monitoring and management of data warehouse will be distributed between HANA, BODS and other tools.

3. No guided process for developing models. Requires lot of self discipline to develop and adhere to development standards and best practices.

4. Hierarchy functionality requires IT support and cannot be maintained by end users like in BW.

5. BW has very easy and versatile functionality for Currency conversion and Units of measure conversion (especially using material alternate unit of measures of SAP ERP). Implementation in HANA is not that flexible.

Kumar

[KulDeepJain · Accepted solution]

Hi Kumar,

Thanks for your crisp feedback on my points, but here I want to add some which give support my points

1. In Standalone HANA architecture we can use DXC method its good one, but let for example if I am using DS as 2LIS_11_HDR/0FI_GL_14. How to create data-targets for this? Because in BW targets DSO/Infocube have there own properties like overwrite/additive. We will again having some standard transformation routines to achieve a particular KPI/characteristics value. Might be in ERP/ECC system we need not to explore about table joins but again we need to understand our data-source functionality. So using just DXC will not solve issue. May be developer will confuse more.
2. In HANA standalone for data acquisition we need separate servers to handle load (SLT/BODS) with separate licensing. BW is mature enough for data acquisition with SAP/non SAP systems.
3. As of now RDS is not available for all modules and also might be not fully mature. (Not fully sure)
4. In BW extractors for any enhancement mostly we need again do full load, But I hope same thing will be there in standalone HANA also. in contrary please give explanation.
5. BODS is flexible as ETL with non SAP flavor, but in SAP BW also we are having DB/UD connect, web-services, ODP, smart data access etc. etc.
6. Standalone HANA give better flexibility to built any application and model, but same can be consume in BW via new info-providers like composite/transient providers... any how vice-versa is also possible.

As per you, It requires full on self discipline while data molding to achieve the best HANA performance, Full agree with you.

Will happy if I will get more distinction between Standalone HANA or BW HANA

KD Jain

[FCI · Accepted solution]

Hi KD,

You raised an interesting topic. Here is my 2 cents.

Concerning the point 4, this is not exactly true. As the replication replicates the whole content of a table (all records with all fields), the need to add a new field is nearly null. This is not the case in the BW case.

And adding a new field to a Hana view (virtual) doesn't need any load at all.

Regards,

Frederic.

[FCI · Accepted solution]

And concerning the point 8 (virtual cube), the Smart Data Access functionnality (SDA) actually allows building in a Hana system a  virtual access to tables or views from other systems (Hana or not).

Regards,

Frederic

[former_member182302 · Accepted solution]

Yes fedric, It allows you to access the tables from other source systems directly into HANA

Regards,

Krishna Tangudu