Skip to Content

Today I had a very interesting discussion at lunchtime about the future of Infocubes and current state of BI and BO integration with two of my colleagues, Michael and Steve. Please read the dialogue:

Michael: “I read today an interesting article about SAP The BW – HANA Relationship. He added, that this is a technology known for years and was third-party companies already do. I read the article, but then I sudenly asked myself, what does this mean for Infocubes?”

Me: “What do you mean exactly?”

Michael: “Infocubes are saved in a star-scheme on the database. With In-Memory technology the star-scheme gets obsolet and is replaced by a flat table structure,  data access is organized by columns and indizes. Why should you think of a star scheme in modeling and take care how you order your characteristics in an Infocube?”

Me: “You’re right. Number of key figures and characteristics in a cube and how they are ordered loses importance. In-memory doesn’t care if it is a flat table or a star scheme what you want to store. I learned at SAP training how to model a cube, how to create dimensions, paying attention to line items. All this gets obsolete with In-Memory.”

Steve: “Yes, I expect that SAP will soon dismiss the idea of a star scheme. Also in terms of Large Scalable Architecture, just think of DataStoreObjects. You won’t need cubes.”

Me: “In my opinion Cubes and aggregates will serve only as fallback, if something is wrong with In-Memory hardware because of any failures. Star scheme is important only for small and mid-size companies, which can’t afford In-Memory technology.”

Michael: “Hardware prices will drastically going down and In-Memory will be affordable for small and mid-size companies in 5 to 10 years. My expectation is that in 5 to 10 years, SAP will dismiss the star scheme for Infocubes!”

Me: “You might be right for this timeframe. In regard of modeling, new BO Data Services come into my mind. With this tools you don’t even need to take care about ETL and writing complex transformations. Data services have a graphical user interface, are easy to lean and fast in data transformations. I expect, Data Services will make it easier to integrate non-SAP systems into SAP BW. You don’t need to write hard-coded ABAPs any longer.”

Steve: “Right, until now, SAP BW is the best warehouse solution, if you have many SAP systems or are a SAP minded company. But there are other customers outside, which run their warehouses on Oracle, Microsoft and IBM. They have sometimes a multi-layered BW architecture, where data from a local SAP BW will be loaded into a global operating Microsoft data warehouse for example.”

Me: “Yes, for these companies BO Data Services is a great tool. And for SAP sales it’s a good possibility to win new customers. That was the idea with acquisition of Business Objects in 2008 – win new customers and making SAP BW attractive for non-SAP customers. By the way, do you remember the reporting problems? How difficult it was to create a customer report with BEx in customers corporate design?”

Michael: “In fact, this has become much easier now. You can choose from Crystal Reort,s WebIntelligence or Dashboard Design. But still you have some landscape issues. How many BO servers, how many BW servers reagarding transports? What about portal server and web server? This landscape is getting very fast very complex and not every company can handle this.”

Me: “As far as I know, with the new SAP BO Business Intelligence 4.0 integration gets even tighter. Let’s watch the video of new features in SAP BO BI 4.0. In my opinion, the tools will grow together over the next two years and known problems today will fade away. BICS as interface gets more and more important. It promises fast access to BI data, replacing MDX interface and queries.”

Steve: “In fact, I have never thought of that. But, yes, you’re right. BICS replacing MDX could be an issue for many third-party tools, which rely on the MDX interface. SAP will enhance the BICS and make it really fast, accessing BI data. Tools using MDX will loose performance and thus acceptance. This can change mind of some customers to question their existing front end strategy. Maybe some of them will change to pure SAP BO front ends, with all migration issues which may arise.”  

Me: “BI 4.0 and SAP BW 7.3 with tight Data Services integration starts to change the world, how you think of BI today and how you have to think about it in 2 years or even less. All inventions together lead to a paradigm change – by customers and partners and how may be BI projects will be done in the future.”

With that, we stopped our discussion and went back to work.

I think, SAP BO BI 4.0 and SAP BW 7.3 ill really change SAP BI in future. I can’t await to get hands on on both tools and try them out! 

To report this post you need to login first.

8 Comments

You must be Logged on to comment or reply to a post.

  1. Sridhar Gande
    I see some exciting comments here about the future of the SAP BI Solutions, I am eagerly waiting to get my hands on to the new tools from SAP.

    Thanks for this blog

    (0) 
  2. Patrick Delage
    I cannot disagree on all except this,

    “Michael: “Infocubes are saved … by a flat table structure,  data access is organized by columns and indizes. Why should you think of a star scheme in modeling and take care how you order your characteristics in an Infocube?”

    When I read this, It’s like the only purpose of Star schema is speed, which I don’t think. It’s a manner to simplify information too, clean the data (I know it should be done before but in reel life…) and also a way to understand data for IT from my experience .

    (0) 
    1. Alistair McEwan
      Because then I need to model delta mechanisms, consistency constraints, table layouts, cube layouts, security, currency & unit conversion, fiscal year calendars, … on my own. Very tedious, extremely high TCO.

      If you do a data mart then BI 4.0 on an arbitrary RDBMS might be ok but not a data warehouse.

      (0) 
  3. Santiago Reig
    Hi Jürgen,

    Nice blog !

    The problem with all this trends of in-memory is the fact that we think of this technology with our head still thinking of the data volumes we used to handle with star schema on top a traditional DB schema !
    I don’t think that with in-memory we will still be able to hold data volumes to the figures we’ve handling until now… I still believe we have arguments to deny massive data volumes into our BI systems because in-memory is not yet an industry standard, but I’m pretty sure that once our BI systems sit upon in-memory techonlogy nobody will accept our arguments when somebody requests to load every single record of operational & financial systems into BI !
    Then a single huge ODS like data model won’t work !

    It was a bit frustrating to understand current Roadmap from SAP and see that in HANA 1.0 there is no support for BW, forcing us to go for BIA (or BWA as it is being called now…) to then go to HANA, once it reaches 1.5 and provides support to BW as well.

    We still lack quality tech info from SAP to fully understand where we’re going and how, but I believe it won’t be so easy to get rid of such a wonderful piece of technology as infocubes based on snowflake schemas !

    Regards !

    Santiago J. Reig

    (0) 
  4. Paul Hawking
    I have asked a similar question to many SAP gurus and only got vague answers.  If the hype matches the reality then the distinctions between OLTP and OLAP become blurred including the supporting structures for OLAP.  These concepts were developed based on the potential of the technology available at the time.
    (0) 

Leave a Reply