Thursday, June 23, 2011

My experience Understanding SAP HANA


I have been trying to understand what SAP HANA Means And what difference it has with Oracle Exadata. Here are my two cents in the controversy.

What is HANA?  (Information Giving by SAP)

SAP HANA is a preconfigured out of the box Appliance

· In-Memory software bundled with hardware delivered from the hardware partner (HP, IBM, CISCO, Fujitsu)

· In-Memory Computing Engine (“IMCE”)

· Tools for data modeling, data and life cycle management, security, operations, etc.

· Real-time Data replication via Sybase Replication Server

· Support for multiple interfaces

· Content packages (Extractors and Data Models) introduced over time

Did you get it????? Mmm To be honest I don't,,, Well in others words HANA is a bunch or bundle of current SAP technologies that include:

-> ETL( Sybase replication Server, Business Object Data Services), Data Base, Modeling Tools, Repository Exchange (ICE ), BI Reporting Tools….

Mmmm a little confuse for me, given that I’m not a SAP expert… Now the question is, What if have to present the same scenario with Oracle technology . Well In Oracle Technology HANA would be equal to:

-> Oracle Exadata,Oracle Database,  Oracle Golden Gate, Oracle AIA, Oracle BPEL And/or SOA, and off course The Oracle Business Intelligence.. Well I have to say that Oracle Database include many additional options that  HANA does not. As partitioning, Compression, Real Application Testing.. Those make a huge difference in my humble opinion.  Another point is that if you are using Oracle Apps and Technology we can say that Oracle In memory Database formerly TimesTen + oracle Database could be fit HANA in oracle World, Off course if additionally we add the others tools (Not the database) Don’t you think?

Another Question show up to me.. Will HANA replace Oracle Database and Exadata Opportunities in Current SAP customers. My answer is yes and not.

First, HANA does not support the current SAP Business Warehouse (BW) , but reading its roadmap they will . Currently a customer who implement HANA today will have SAP BW and HANA at the same time.. Not worthy on my point of view. For Now HANA has In memory capability and Sybase replication server. That means the data is replicated and stored in memory , so then BW ETL tool is connected to HANA, making the extraction fast and the ECC which is connected to HANA sends the data on real-time or nearest to real time.

Secondly Will HANA Support the SAP Business Application (BA), Yes but is not available yet.. They are planning to support BW first and later in a future the BA.

Now If you have SAP BW and performances issues.. What is the best approach to fix them?

Option 1. HANA (Another Server, a new implementation , a lot new licenses and and more more money)

Option 2. EXADATA (Replace its current Oracle DB Server, it means move your current Oracle Database Supporting the SAP BW to EXADATA Machine, some additional licenses and no implementation project, and yes some additional money.

In my Opinion Option 2 give us better ROI and can be in production faster. 

Well As soon I got more inputs I will share with the community.


  1. Some interesting things here and some thigns I disagree with.

    First, compression and partioning on HANA are implicit. It's a columnar database so it's compressed 10:1 compared to Oracle, and partioning isn't relevant for in-Memory worloads.

    Those customers that are implementing HANA 1.0 are implementing it because of specific business needs for which super fast analytics are worthwhile. So having both BW and HANA is perfectly valid in this context.

    If you have BW performance problems then today the best thing to do is to implement the Business Warehouse Accelerator, which is an in-Memory addon for BW.

    In Q4 we expect to see HANA 1.0 SP03, which will support NetWeaver BW as a platform, and should (if the architecture is anything to go by) outperform Exadata. We will see.

  2. Thanks for your comment, the only I would like to know is about compression. In exadata his Hybrid columnar Compression meaning By Column and Row and the same time, reach sometimes more than 50x in compression.

    Which one is better?.. mmm I do not.. Could be a god idea find some one who have tested both of them.

    Other personal oppinion is that HANA 1.0 looks more like Oracle Exalogic than Exadata.. And in the future HANA will be as Exalogic and Exadata Combine.. So much interesting ..