Home > Blog > Blog > 2012 > June

You have questions around BW on SAP HANA? You are wondering how startups implement SAP HANA? And you’re familiar with Twitter?

Since the last SAPChat around HANA we’ve seen a lot of questions around BW and HANA on Twitter and we want to give you first hand answers from our exclusive Expert Panel: Steve Lucas, SAP Mentors Aiaz Kazi and Vijay Vijayasankar and Lou Leporace from NextPrinciples representing the startups adopting HANA. This is your chance to ask our panel of experts questions by using the #SAPChat hashtag and we will answer them for one hour on Thursday, July 12 at 9 am PT. The tweetchat will be hosted by SAP’s social media marketing team.

You haven’t heard about recent discussions on BW on HANA? Or how Startups and other companies build applications on top of SAP HANA?

SAP works with startups and hosted two startup forums where startups present their solution and how they can benefit by SAP HANA. In addition, there have been many questions on whether SAP HANA will replace BW and we’ve had several blog posts addressing these questions and concerns. However, what is better than asking the experts directly? Therefore we will cover all these topics in the tweetchat and explain how to connect the dots between the HANA Platform, SAP NetWeaver BW and the adoption from startups and other third parties developing applications on top of HANA.

You don’t know what a tweetchat is?

Think of it as a Q&A where the speakers type out their thoughts in 140 characters or less. And instead of having a one-way discussion, you will help to drive the conversation. And during #SAPChat, rather than using a dial-in number, you follow with your preferred Twitter app like Tweetdeck, Hootsuite or even TweetChat.com. To participate in the tweetchat, simply log in to Twitter and submit your questions with the hashtag #SAPCHAT. Steve, Aiaz, Vijay and Lou will be monitoring the timeline live and responding with their accounts @nstevenlucas, @aiazkazi, @louleporace and @vijayasankarv

When will this take place?

On July 12, please join us from 9:00-10:00 A.M. PT (12:00 P.M. ET) using the #SAPChat hashtag. During this hour, they’ll be monitoring the timeline live and responding with their accounts @nstevenlucas, @aiazkazi, @louleporace and @vijayasankarv. The chat will be moderated by @SAP. As SAP HANA social media manager I’m really excited to participate and to have Steve, Aiaz, Lou and Vijay on board. Are you ready? Think about your questions! We’ll be waiting for you on the 12th!

Last week was very exciting for the whole SAP HANA Team, but also for all the HANA followers and users out there! SAP HANA turned 1 and we celebrated all over the world in several SAP locations, such as Germany, Canada, Brazil and Palo Alto. As social media manager for SAP HANA, I had lots of fun to engage with you via our social channels and make sure you enjoy our big event. Now I’m going to take you once around the world with impressions from SAP HANA’s first Birthday.

The birthday started with Vishal Sikka taking the stage and talking about SAP HANA and the key achievements within just one year after being announced generally available. And these are really impressive! We are talking about 358 customers, 159 implementations, +65,000 end users and 16 customers in the 10k+ performance improvement club. One key announcement made by Vishal Sikka during this special day was the development of a Distinguished Engineer Program which will be a certification of "hands on SAP HANA experience". Read more in David Hull’s blog.

     Our colleagues in Newtown Square, Sao Paulo, Dublin and Vancouver.


And there were not only 200+ people in Palo Alto, including SAP Mentor Vijay Vijayasankar (Thank you again for taking time and celebrating with us in person), but more than 1,000 people online and in other SAP locations. Having so many people celebrating at the same time was amazing. Our colleagues in Walldorf, Germany waited one more day to celebrate and experience Vishal live talking about SAP HANA’s first Birthday.




If you want to see SAP HANA's milestones, look at our SAP HANA timeline:




Now you’ve seen a quick recap from SAP HANA’s first birthday, but if you really want to experience Vishal Sikka’s excitement about HANA turning 1 you should watch the replay. We're having a lot of SAP HANA Birthday activities planned throughout the summer, so make sure to follow along and mark your calendars for an upcoming HANA #SAPChat on July 12. More information to follow in the next few days.


On June 14th, 2012 SAP Labs in Palo Alto played host to 35 innovative startups at the next installment of the SAP Startup Forum. The event was a full day of learning and discussion, centered around the theme of HANA on mobile applications. 15 of the 35 startups presented their ideas and product demos to the audience, showcasing how their solutions could benefit from implementing HANA. The presentations were interspersed with brief “101” sessions on SAP technologies, programs, and platforms which the startups were curious to learn more about.


After lunch, it was time for a deep dive into four breakout sessions. In these sessions, participants had an opportunity to dive deep into SAP HANA, SAP Ventures, SAP Mobility, and the Startup Focus Program itself. These breakouts gave the startups a place to discuss any concerns they may have had with SAP executives who could answer any of their questions.


The Startup Forum finished off with a mixer held in the cafeteria. The startups set up small booths where they highlighted their products and solutions. SAP employees were invited to hear the startups’ pitches and “invest” fake SAP euros in the company they felt had the highest chance of success. Over a billion “Euros” were in play and the “winning”startup, the “People’s Choice,”  received over 150 million Euros, and the SAP employees who invested were entered in a drawing for amazing prizes. (Include pictures) Congratulations to the four startups who went home with awards:

  • The “Best Mobile on Big Data Solution” went to Mobile Intelligence Solutions. MIS provides a market leading analytics solution to mobile carriers that delivers strategic, actionable insights on mobile data activities.
  • The “Most Innovative Mobile Solution” was won by Liquid Analytics, a company which integrates enterprise technology with mobile applications.
  • The “Best 140 Character Pitch” was taken home by iSky, a startup that predicts customer actions to provide an enhanced customer experience.
  • The winner of the “People’s Choice” award was Altia Systems. Their solution makes it possible to create panoramic videos on mobile devices, augmented with individualized data on people, places and things.


All in all, the second Startup Forum was a great success. Startups and SAP employees alike were able to learn more about one another in a setting that encouraged networking and open conversation, and a few lucky winners walked away with awards and prizes. Look out for the next Startup Forum in your area!

Two major benefits achieved by migrating SAP NetWeaver BW to the SAP HANA database are (1) the dramatic improvement in data load times for Data Store Objects and (2) the significantly boosted performance while reporting from them. With the latest content releases found in SAP NetWeaver 7.30 BI Content 7.37 SP 01 or SAP NetWeaver 7.31 BI Content 7.47 SP 01, a large fraction (about 2/3) of DataStore Objects that were delivered so far have now been prepared for HANA-optimization.


When you copy these HANA-prepared DataStore Objects (DSO) from the delivered to the active version, they will be created automatically as SAP HANA-optimized DataStore Object side-stepping the manual conversion step from standard to HANA-optimized DSO. The delivery of DataStore Objects prepared for HANA-optimization marks the next significant step towards total cost of administration reduction after the migration of all relevant data flows to SAP BW 7.x technology. This is one important prerequisite for HANA-optimization of involved DataStore Objects (migrated data flows were delivered with SAP NetWeaver 7.30 BI Content 7.36 SP 02 or SAP NetWeaver 7.31 BI Content 7.46 SP 02). If DataStore Objects contain data, you cannot avoid the manual conversion step to HANA-optimization. Therefore, you best benefit from the HANA-prepared DataStore Objects when you copy new data flows from the delivered to the active version in your BW system powered by SAP HANA.


In addition, you can benefit from automatic creation of HANA-optimized DataStore Objects during Content activation if you are using (1) SAP HANA database as of release SAP HANA 1.0 Support Package 03 and (2) SAP NetWeaver BW 7.3, Support Package Stack 07 or SAP NetWeaver BW 7.3, including Enhancement Package 1, Support Package Stack 04.


Find the complete list of all 1025 HANA-optimized DataStore Objects attached to note 1708668.

Without going into a history of inaccurate and inconsistent statements about SAP HANA made by Oracle and Larry Ellison himself, I just want to use this opportunity to set the record straight on something that they continue to cause confusion about. In the Oracle earnings call on June 20, 2012, this is what Larry had to say (http://seekingalpha.com/article/670061-oracle-management-discusses-q4-2012-results-earnings-call-transcript) about Exalytics and SAP HANA:

“….And Exalytics, our in-memory analytics competitor to SAP's HANA, is ramping faster …..  and we're not seeing much competition from HANA.”


Here is a summary of what the market literature tells us about Oracle’s Exalytics:

It is meant to be an analytics appliance constructed by adding multiple technologies and duplicate engines into one box for different use (e.g., Essbase for planning, and TimesTen for caching, and other technologies such as OBIEE). Also, it is clear from public documentation that there is a 1 TB limit on Oracle Exalytics, and from all indications it appears that only about 400 GB of this is actually available for in-memory data since a significant portion is to be used for working memory for the components (e.g., Essbase, TimesTen, OBIEE) that have been put together. It is clear that data needs to be curated (ranging from pre-selection to pre-aggregation) before it can be used for meaningful analytics operations.


So yes, indeed it does provide a limited in-memory experience, and, yes, if you are using pre-fabricated data then it does provide a fast response. That’s where the similarities end. SAP HANA is more than an analytics engine. SAP HANA provides a full-blown in-memory experience that is unrestricted by how much or what data you use. It can be used for both analytical and transactional data. There is no data caching involved and there is no requirement for any tuning, or for pre-aggregating or otherwise pre-fabricating the data. As the growing number of apps and solutions on SAP HANA, and the success of BW on SAP HANA prove, this is much more than a mere analytics appliance.


Hence, while Exalytics can seek to compete on a limited number of aspects, it is by no means a “competitor” for SAP HANA. For Oracle to present something that is a true competitor for HANA, they will need to find something that can scale out to 500TB of raw data stored in main memory and deliver real-time results without manipulating the data first!


While all these limitations of Exalytics (vis-à-vis SAP HANA) might not hurt a customer in certain scenarios, they are certainly a drag in many other cases. Hence Exalytics as it stands today can never address the comprehensive scenarios that SAP HANA can.


One solid measure of a solution's true worth is the growth of its customer base. Since its introduction a year ago, SAP HANA has 354 customers, and counting!

P.S. You can follow my other blog posts at: Café Innovation Blog on the SAP Community Network (SCN)

Many of the 13,000+ SAP NetWeaver Business Warehouse (BW) customers worldwide have adopted SAP BusinessObjects BI solutions as the new face of BW – or are in the process of doing so – in order to take their business intelligence to the next level data exploration and insight. By doing so, they enable fact-based decision making across the entire organization above and beyond the usual and rather focused SAP Business Explorer (BEx) user base.


As they now adopt SAP HANA as their database of choice to power their BW data warehouse they have even more value from their investments with accelerated BI. Let’s see how.


Let me start by explaining that the full suite of SAP BusinessObjects BI solutions (BI 4.0) comes with a unique integration to BW and SAP landscapes in general with multiple optimizations aimed at delivering very tangible value to line of business users and IT departments.


All this value remains available to customers as they move their BW applications on top of HANA. This move is transparent to BI 4.0 and BI users simply enjoy a performance boost for their existing analyses, reports, and dashboards. In addition, BI 4.0 as the face of BW powered by HANA will also unleash new self-service BI scenarios: all SAP BusinessObjects Explorer users will explore their business at the speed of thought without boundaries to the BW data that will be exposed to them in HANA. For example sales representatives on the road to visit customers will have unlimited and immediate access on-the-fly to all the information they need to support live customer discussions.


In the longer term more SAP BusinessObjects BI functionality will benefit further from the power of HANA. Whether it’s pure set slice-and-dice or more hierarchical drill-down, whether you’re doing it an Excel metaphor, a Web browser metaphor, or on a tablet metaphor, the power of HANA will surface through BW to accelerate and innovate your BI.


Need to learn more? Go the dedicated pages on the SAP Community Network: http://scn.sap.com/community/bi-platform/businessobjects-bi-for-sap

Steve Lucas wrote a great blog this week entitled Does SAP HANA Replace BW? (Hint: No.). Given he covered at a high level a few of the dozens of possible scenarios applicable to BW and HANA, I thought I would address a few more. That and at the end of it he mentioned there would be a Part 2…so I immediately asked him if I could help him write it!


Sometimes I feel like people forgot why they implemented a Data Warehouse in the first place and therefore don’t see much value in it…which is concerning and a great place to start.


Q: Why did or might customers want to implement SAP BW?


By far the majority of SAP BW installations are for SAP ERP customers, many of whom wanted an Operational Data Store. We often call SAP an Enterprise Data Warehouse (EDW) but it is rarely used as one. The reasons for implementing it, however, are the same:


  • History. Retain historic information that you want to remove from your ERP environment.
  • Persistence. Integrate information from disparate sources, store it in one place and present a single view of the truth.
  • Performance. Remove load from ERP environments and aggregate data for good performance.
  • Dimensional Conformance. Retain same semantic meaning across process areas – of customer, supplier, product etc.


My supposition is that the reasons above are why customers deployed SAP BW. They may not understand or appreciate all those points, which is concerning! This brings to mind a few questions, comments I have about the legacy of BW and its future.


What’s the issue people have with SAP BW? It’s all about performance!


Steve gets into great detail in Part 1 of the blog, and the essence of it for me is that traditional RDBMS like Oracle or IBM DB2 haven’t provided what is needed. When organizations implemented SAP BW back in the early 2000s, the data volumes and complexity allowed simple Operational Data Warehouses to perform reasonably well when run on an OLAP database like Oracle.


We’ve had two issues working ‘against’ BW over the last decade:

1) Massive data volumes stored in BW, way beyond initial design. The BW product has gained new features and at the same time, data volumes ballooned and the use cases for BW increased dramatically, so organizations had packed much more information into BW than originally planned/modeled.

2) Performance expectations have moved from hours or minutes to real time. The rise of Google and the iPhone means that people expect information at their fingertips, without a wait.


Granted, the Business Warehouse Accelerator (BIA/BWA) goes some way to resolve the response times, but regardless, the problems of agility, adoption and load times remain, and BW has become bogged down by its own success.


What if I just use BW for one area? Will that help performance?


If this is the case, then you aren’t using BW as it was really intended. While that’s perfectly acceptable, there may be better options for you to consider. If you don’t need to derive the four main benefits of BW I mention above, then you have the complexity of a Data Warehouse without the benefits. Not good.


In this case a good approach would be to use HANA Enterprise without BW, which is an excellent database and data mart, using either SAP’s Direct eXtractor Connection (DXC) or System Landscape Transformation technologies as the ETL layer.


But watch out if you are in fact an extensive BW user and you decide to take this approach. Your first HANA standalone Data Mart project on HANA Enterprise will work great. However, when you come to a “second” project, you will end up with multiple Data Marts rather than an integrated solution – losing in the process a bunch of benefits that BW gave you, which you forgot you wanted.


It’s worth remembering as Thomas Zurek points out, “It's not an either-or but an and”. You can have HANA Enterprise as a Data Mart, and BW on HANA as an Operational Data Warehouse. Have your cake and eat it!


What if someone has already invested in BWA?


Big shout out to Emma Moss for asking me this question. If a company has deployed BWA then they are already in a great position because they have experienced first-hand the benefits that in-memory computing can bring – to the reporting layer of SAP BW. (BWA/BIA is an In-Memory system, albeit different than HANA)


Let’s say you own BWA and Oracle, and want to migrate to a “HANA only” scenario. If you do, there are actually quite a few benefits:


  • Simplification into one single database – not one Oracle DB and a bunch of BWA blades.
  • The same great in-memory benefits to your Data Warehouse development and data loads.
  • Simplification of data models and cutting out of redundant layers. (DB Indexes, Aggregates, BWA Indexes, InfoCubes etc.)
  • Reuse of rather expensive mainframe BW hardware and commodity BWA blades into your IT landscape.


All that and SAP will typically provide a credit of your BWA purchase price towards your HANA purchase. It’s worth talking to SAP about that.


What if we are considering getting rid of BW because it’s too slow?


I talk to a lot of customers who are frustrated with BW and some of them have good reason to be frustrated. However, there is one key question to bear in mind if you are considering moving away from BW:


First and foremost…What would you move to? Steve puts some suggestions of what might be on your mind: Rapid Marts, Custom Build, etc. but they are fraught with challenges (see my points above). You have the same problem with replicating all the complex business transformations in another EDW like Cognos. Besides, are you sure that moving from BW to another traditional EDW platform is going to make things better?


What’s more – organizations spend millions, tens of millions or even hundreds of millions building out an EDW. Building out a new one will cost you the same again. Business Processes don’t magically simplify themselves because it’s 2012 vs. 2002.


BEFORE you decide that BW is “not right” for you, consider that swapping out your Oracle DB underneath and existing BW system for SAP HANA will cost a relatively modest amount of money compared to building out a new EDW, and the cost can be determined to a very high degree of accuracy early on. Once you get to BW on HANA – many of the objectives you had in mind…e.g. speed, performance etc. will be right there for you.


To add to that, you don’t have to undergo business change – the reporting tools can stay the same if you want them to – and you don’t have to have a long, expensive change freeze.


Oh, one more thing. SAP is so confident that you’ll be thrilled with BW running on HANA, they have a deal on the table: if you’re not happy with SAP HANA in 18 months, SAP will provide you with a license credit for another product.


Why do I need BW at all if I have DXC?


Thanks to Jan van Ansem for this question. You are absolutely right that the DXC technology – a mix of the ERP/BW Business Content Extractors and the Data Services product – allows you to use SAP Business Content and load that into almost any database. This is useful in some non-BW on HANA scenarios:


1)   When/If you want a data mart. As with the scenario above, if all you wanted was a data mart anyhow, DXC gives you options of where you want to put it. I recommend SAP HANA Enterprise because it is awesomely fast and flexible, plus is comes with some RDS scenarios that include HANA content, but you could potentially use any other database including Sybase IQ, or a 3rd party EDW.

2)   When you have an established EDW. If you have an established EDW and you only want a bit of ERP data then it may make sense to move this data into that EDW using DXC.


However if what you wanted was an EDW, then BW on HANA makes a whole lot of sense as that EDW – not any other platform. Why is this?


1)   BW is the only EDW that provides the “other half” of Business Content. Whilst the SAP ERP Business Content Extractors provide one half of the business content (including the handling of delta changes), if you use any other EDW platform including Sybase IQ then you have to rebuild all the business transformations, structures, cubes, queries etc. that exist for free in BW. Don’t underestimate how much work this entails.

2)   BW takes some serious heavy lifting away. Features like stock pointers, time-dependent master data, virtual cubes, MDX and SAP BusinessObjects integration etc. make BW a very compelling platform. Before you discount BW, make sure you know the effort required in modeling these in another EDW.

3)   BW on HANA gives you what you wanted in the first place. When we combine BW with the power of the HANA Database, you get what you really needed: a powerful modeling platform with the agility and performance that you wanted.


Given Jan asked me this question, let’s address his valid/typical BW pain-points and discuss how BW on HANA might address them:


1)   Point: BW can present complicated, long running load processes which are difficult to manage.


Counterpoint: BW on HANA allows you to collapse layers. With BW on HANA you lose Databases, Indexes, Aggregates, BWA indices and Dimension Tables on day one. This simplifies things, especially on BW 7.3 which has much better administration tools for loading. Plus, in the future, you can look at removing cubes altogether – depending on your scenario – and creating logical cubes rather than physical instantiations.


2)   Point: A delay of one day is possible between when data is loaded in the transactional system and the data being available in BW reporting.


Counterpoint: BW on HANA allows you to combine real-time and batch data. You can load data into HANA using SLT in real-time and combine this with BW master-data if you want to, using Transient Provider technology. In addition, the reduced number of layers means that you can load more often.


3)   Point: Business users complain about the reporting tools in BW. (BEx Analyzer, Query Designer, Web Application Designer)


Counterpoint: SAP is investing in the front-end and much of this is available today. The SAP BI4 platform – especially in the new FP3 release – is highly integrated for SAP HANA. What’s more there are two new awesome tools – SAP Visual Intelligence (Codenamed HILO), which is a desktop analysis tool and SAP Predictive Analysis, which is also a desktop tool but a business analyst-friendly SAS/SPSS competitor. And there is SAP Analysis Edition for Application Design (known affectionately as project Zen) coming later in the year, which replaces the Web Application Designer and will be mobile-friendly.

4)   Point: The reality of poor performing reports.


Counterpoint: Hopefully I don’t need to tell you why BW on HANA fixes this! But, SAP is investing substantially into the future of BusinessObjects technology to make it integrate much more deeply with HANA – for example moving the front-end technologies of the BI platform and the Data Services ETL engine in-memory.


If one comes to the conclusion that BW on HANA is a good option, one will naturally have additional questions.


What will happen to ABAP in BW?


Thanks to Vivek for this question. The short version is: HANA is a database that replaces Oracle for the RDBMS on BW and the application layer works just like it did before (but better!), and that includes your existing ABAP.


In addition, stuff like the OLAP Engine, DSO Activation and Planning functions are pushed into the in-Memory HANA layer down from the ABAP engine. This makes them really fast.


Regular ABAP continues to run, against the HANA RDBMS. Or you can optimise it for HANA if you like, including using the powerful SQLScript Stored Procedure language run from ABAP – or alternatively just by rewriting your custom ABAP in transformations, DTPs, etc. to take advantage of the way that SAP HANA can process set-based SQL statements much faster than loops.


My advice: find slow-running custom ABAP and optimize it for SAP HANA, and leave the rest as it is. This way you get the maximum benefit for minimum work!




If you are an SAP ERP customer, need an analytics solution and asking the question: “Should I implement or move to SAP BW on HANA?” then I think the answer should be: Yes, unless you are one of the special cases above where you don’t need it.


Hopefully this Part 2 of the blog will necessitate Part 3 – but that’s up to you as to whether there are still scenarios and questions left unanswered!

I have been part of the SAP HANA marketing team since early 2011 and was incredibly proud to be part of the celebrations honoring the first anniversary of SAP HANA becoming generally available. The event inspired me to consider the impact that SAP HANA can have on us as individuals and our daily lives


Transforming Medicine

Our health is something we value more when we get sick. Almost all of us have some personal experience or know a dear loved one go thru the agony of not knowing if they have a serious illness such as cancer. Imagine if your healthcare provider did health3.jpgnot prescribe treatment to you based on how the majority have responded to your condition but treatment that is based on your individual DNA. Currently, it takes two days to find differences in genome data between having cancer and not having it. Genome Analysis and Personalized patient treatment are two use cases that describe how organizations are already leveraging SAP HANA. 

Health insurers, such as AOK in Germany plan to analyze the tremendous amounts of medical data from 24 million of patients, over 370 million outpatient treatment cases, 1.5 billion diagnoses to recognize potential health risks, create preventive care programs and appropriately respond to identified risks ahead of time.Hence, the impact to me as patient is that advanced medical treatment becomes much more accessible while at the same time every interaction with the medical profession becomes much more personalized and effective.


Transforming Travelling

As a traveller the airport security screening process though very important is also time-consuming and stressful. Especially travel.jpgduring peak holiday season like now.  Imagine if airport security systems could use sophisticated screening mechanisms throughout the airport to automatically detect and react to threats in real-time instead of treating every single traveller as worst offenders. More than 30 million travellers in India travel daily on the Indian Railway network. On popular routes in peak season the tickets are sold out within hours of opening, even though bookings open 120 days in advance from date of travel. SAP HANA has the potential to help millions of Indians plan their travel experience


Transforming Shopping

How often have you just bought a large cost luxury item at a store only to receive a promotional offer for the exact item a few retail.jpgdays later? Imagine if while you are in the store the retailer could predict and offer you an incentive instead; based on your individual profile and past buying behavior. Real-time insight into customer behavior is a use case that highlights how a large Japanese electronic retailer that has about twenty domestic stores located at major transport hubs in large metropolitan centers within Japan faced exactly this challenge. How to enable shop floor staff to quickly gain access to specific customer data while the customer was in the store so that they could provide personalized (best match) services. Analyzing POS data with HANA provides the best recommendation to the customer. 


Transforming citizen services

Since the start of the financial crisis, reducing government spend is  becoming a priority. Reductions in government spend usually leads to reduction in quality of services for citizens. However there is public.jpganother option open to public departments: Government agencies are making billions of dollars in improper payments each year and are under pressure to eliminate fraud, waste and abuse in government spending. For example, the US government accountability office estimates that improper payments at the federal level totaled more than $125 billion in 2010.  Governments can fight fraud, waste and abuse with predictive analytics from SAP HANA, thereby substantially reducing improper payments and establishing better processes by identifying and preventing improper government payments before they happen.



The impact of how HANA can impact and transform our lives is only limited by our imagination. Share your examples by commenting on this blog or sharing your HANA use case.





The much anticipated 1st birthday celebration for SAP HANA is here and what she’s been able to achieve in such a short time is truly remarkable. We know that HANA has 350+ customers with over 140 go-lives. We’ve also heard the latest news that HANA clearly enhances BW and will not replace it, and finally - that ERP will run on HANA by year’s end. 


The original Birthday of SAP HANA is June 21, where HANA was announced GA in 2011. However, we are celebrating today, on June 18 2012 all over the world. Clearly there are two questions that are a bit more pressing... Which famous people share HANA’s birthday celebration, and what major events happened on this day?

A quick Google search shows over 200 people; as I sift through this data (much slower than HANA of course), I come to find many Prince’s & Princesses, Writers, Painters & Musicians, along with Athletes, Actors & Admirals (including the original voice of Superman and the original Tarzan). One thing’s for sure, there’s no shortage of influence and clearly HANA falls within the appropriate group – so without further ado, here is my top 5 list of famous birthdays (and since this article is shockingly only meant for entertainment, this list will stay within the entertainment world):

     1. Paul McCartney, 1942 – vital cog in the greatest band ever created, I’d have to create a new word to fully describe the sweet bliss of listening to the Beatles.  The Beatles are the bestselling band in history with estimated sales of over 100 billion units.  Advantage: McCartney over HANA (although her pipe is growing)


     2. Roger Ebert, 1942 – Famous film critic (Siskel & Ebert), his reviews are syndicated in over 200 newspapers.  Ebert has compiled a “best of the year” movie list since 1967 but while he critiques the best, HANA focuses on being the best.  Advantage: HANA over Ebert


     3.  Lou Brock, 1939 – Hall Fame Baseball player known for breaking Ty Cobb’s stolen base record with 938.  Clearly he’s fast, but not the fastest as Rickey Henderson eventually “ran faster” swapping a total of 1,406.  Advantage: HANA over Brock

     4. Bruce Smith, 1963 – Hall of Fame Football player and all-time leader of sacks.  Bruce Smith played for the Buffalo Bills and appeared in the Super Bowl 4 years in a row.  Clearly a man of power and although his record will one day fall, until it does – he still gets the nod.  Advantage: Smith over HANA

     5. Darren “Dizzy” Reed, 1963 – A perfect example of Aristotle’s, “The whole is greater than the sum of its parts”, Dizzy Reed is better known for his contributions as they keyboardist/pianist for the popular rock band, Guns N Roses. Sorry Dizzy, HANA is making a name for herself, not just SAP.  Advantage: HANA over Dizzy Reed


Now let’s see how June 18th stacks up historically:


1682 - William Penn founds Philadelphia, US

1778 - British Redcoats evacuate Philadelphia

1815 - Battle of Waterloo; Napoleon defeated by Wellington & Blucher

1873- Susan B Anthony fined $100 for voting for President

1892 - Macadamia nuts 1st planted in Hawaii

1928 - Amelia Earhart becomes 1st female to fly across Atlantic Ocean

1960 - Real Madrid wins 5th Europe Cup 1

2011 – SAP HANA introduces a new paradigm of real real-time computing.  Dr. Vishal Sikkha states, "General availability is a major milestone for SAP HANA, but this is just the beginning and I look forward to the next milestones with SAP HANA, building breakthrough new applications and building out the SAP HANA application cloud to deliver them."


Clearly a lot of historic events have occurred on this memorable day and SAP HANA has only added to the resume by inking herself into the history books and to Vishal’s point, the future is now!  So join me in saying, “Happy Birthday HANA, your best years are still ahead of you!”



“What does a future – in which the data, knowledge and expertise from all of humanity is instantly available to everyone – look like?"  Check out http://www.aswemaywork.com as we explore one little girl’s story from a sea of possibilities.

Since the general availability of SAP HANA one year ago, there has been tremendous interest and excitement in the technical community regarding the adoption of HANA skills. This groundswell has led SAP to develop and release several offerings in support of those wanting to learn how to support and develop for HANA, the most recent of which is the release of free developer licenses for the HANA platform through Amazon Web Services.


Continuing this trend, SAP is proud to offer support to a new community-driven organization that aims to promote and encourage HANA technical expertise, called the SAP HANA Distinguished Engineer Program.


The Distinguished Engineer Program will serve to bring together hands-on HANA technical experts in the marketplace in order to more easily facilitate the sharing of knowledge, lessons learned, success stories, cross-training, content and best practices. In addition to this, SAP will support this organization by providing HANA briefings, development and roadmap updates, access to resources, and joint coordination of activities at conferences and other virtual and in-person technical events. SAP will also solicit feedback from these experts regarding real-world HANA implementation experiences to use as input to guide development and product management for future patches and releases.


The leadership of the Distinguished Engineer Program consists of some of the foremost HANA Influencers and Mentors in the marketplace, including Vijay Vijayasankar, Head of SAP Forward Engineering for IBM Global Business Services; John Appleby, Head of Business Analytics & Technology for Bluefin Solutions; Harald Reiter, Senior Manager with Deloitte Consulting; and Jon Reed, Independent Analyst, Blogger, and specialist in SAP career planning and skills development. Membership criteria for those interested in joining the HANA Distinguished Engineer Program will be established by the program leadership soon and published openly.


SAP is excited at the prospect of this new and unique approach toward community-driven technical excellence, and we look forward to working closely with HANA Distinguished Engineers as we strive together to support innovative customer solutions.

*To read an update on this blog, please click here: Does SAP HANA Replace BW? (Hint: Still, No.)

Well – I can’t say I am surprised by all the debate on Twitter about SAP HANA potentially replacing BW. I wrote this blog in response to about 50 tweets I saw on Friday about the topic. I want to thank Vijay Vijayasankar from IBM for his contributions! I plan on covering four scenarios in this blog that will hopefully help guide people. Those scenarios are:


1. I have SAP ECC and BW, what do I do?

2. I have SAP ECC and may want BW, what do I do?

3. I have SAP ECC, but someone told me SAP BusinessObjects Rapid Marts are better, what do I do?

4. I have SAP ECC and someone told me they will design a data warehouse for me from scratch, what do I do?


The easiest place to start answering these questions is by defining BW first. The reason being that it’s fairly obvious that not everyone defines “BW” the same way. Some people don’t even call it BW, but rather SAP BIW or BI. In defining BW, I will first quote the source of all truth, Wikipedia: “SAP NetWeaver Business Warehouse (SAP NetWeaver BW) is the name of the Business Intelligence, analytical, reporting and Data Warehousing solution produced by SAP AG. It was originally named SAP BIW (Business Information Warehouse), then abbreviated to SAP BW, but is now known as "SAP BI" at the end-user level. In contrast, "BW" is still used to describe the underlying Data Warehouse Area and Accelerator components. It is often used by companies who run their business on SAP's operational systems.”


The posting on Wikipedia further goes on to state that “SAP’s BW solution has a pervasively employed data warehouse and contains a large number of pre-defined business content (pieces) in the form of InfoCubes, Info Objects, authorization roles, and queries. This allows the ability to leverage SAP's experience and to reduce implementation cycles. The business content can be modified to meet an organization's specific requirements; however, this requires a longer process of customization of the pre-defined elements.”


Finally, the posting states that BW has several components, including, but not limited to:

  • An optimized Extraction, Transformation and Load (ETL) layer for moving data around.
  • A Data Warehouse Area that stores data in a star schema design.
  • A reporting component for consuming information from the data warehouse.
  • Planning and analysis to simulate operations like budget calculations.


The point I am trying to make here is not to over-quote Wikipedia. But since I am writing a blog on whether or not HANA replaces BW and there are barely two people on the planet who will define BW as the same thing, a little grounding can’t hurt!


Now that we have the official definition of BW, here’s my definition: It’s a system. A system consisting of a warehouse, a pre-defined business schema, as well as content and related tools for improving analysis of data stored in the SAP Business Suite. Plain and simple. Either all or part of this system has been deployed at over 13,000 customers, it has thousands of third party apps certified by SAP running on top of it, and it delivers reliably (note, I didn’t say quickly). And BW is free. But it’s NOT a database. It works ON top of a database, and that distinction needs to be kept in mind. You also spend money on a slow disk based database to power your BW today and often accelerate it with a hardware solution like Business Warehouse Accelerator (BWA).


Moreover, for the past 15 years, for every module in the Business Suite (not quite, but pretty much most of it), we’ve built logic to convert the 4 character table names in SAP ECC to have "business meaning" via BW content. So the question to contemplate as you read this blog post is: Should one set out to rebuild content that SAP spent 15 years building for you because it's perceived as "difficult"?


As for my definition of SAP HANA: It’s a database. (You can’t use that quote without the rest of the story below...).


Granted HANA is a special KIND of in-memory database that delivers insanely great performance when coupled with systems LIKE BW, but at its most basic level, it’s a database. HANA does not replace BW. It is meant to enhance BW. On the speed note, it’s worth noting that I view one of the major challenges with running BW on a traditional RDBMS as poor reporting performance. With HANA as the database for BW, reports (amongst many things) absolutely perform better. If you can get SAP BusinessObjects (BOBJ) as a front end to BW, even better.



My ENTIRE point: BW is MUCH better on HANA, and coupled with the fact that BW is free, there is a ton of pre-built content for BW A


ND you get instant certified solutions on top of BW-- isn’t it worth at least CONSIDERING it if you don’t own BW? Remember, there is NO plan to sunset BW. If someone has stated that, they are unequivocally wrong. I am a BW plus HANA advocate instead of a BW vs. HANA guy.

I get that there are options to BW. Some customers have acute needs and will find a BusinessObjects Rapid Mart a quick way to move data using Data Services into a much more manageable “store” than BW. But you also will have to build (or, customize, depending on the Rapid Mart) the extracts, transforms, reports, etc. you’ll need in order to be productive. I know SAP is building some new content in HANA natively, like the Operational Reporting RDS. So one might wonder if SAP is "switching gears" in terms of where c


ontent is created and resides. (Again, the answer is no, we haven't switched gears at all.)


I get that some people will say “just design your own warehouse from scratch,” but the same issues apply. It’s classic ‘build vs. buy’ and a BW scenario with HANA or any other database is NO exception.



I am going to cover the four scenarios I promised you, so here they are:


Scenario #1: I have SAP ECC and BW, what do I do?

The short answer: Companies should absolutely leverage HANA as the database for existing BW deployments.


I mentioned earlier that speed of reporting is a big pain in BW without HANA; however, its really about end to end performance – loading, reporting, simple landscape and change management.  Due to parallel loading we dramatically reduce the time of loading data into SAP HANA for BW and it’s a essential for companies that have long nightly jobs (get HANA, spend time with your family rather than babysit these loads!)


Speed of ETL and Reporting is only one aspect of HANA as the database for BW. An equally important - maybe even more important - aspect is simplification. BW has some extra layers that were built solely for performance reasons. Those layers are not needed when HANA works as the database for BW. A leaner BW is more efficient to build functionality in, and manage in a production environment. The Layered Scalable Architecture for BW is evolving to make the best use of the power of HANA as its database. Functionality that used to work in the ABAP layer of BW is now being moved over time into the HANA layer to make the entire model and processing work much more efficient. An example of this is DSO activation that happens in a fraction of the time it takes with a non-HANA based BW system.


Often, many BW customers have SAP Business Warehouse Accelerator to accelerate their slow disk based RDBMS for BW. SAP HANA provides a much simpler landscape reducing TCO and complexity. It reduces your hardware footprint dramatically - e.g. to accelerate 5TB of BW data, you would need 21 blades in BWA  vs. 1 HANA server with the added benefit of no third party database since HANA is the single persistent database. It’s a no brainer!


Change management is relatively easier with the choice to build more logical layers instead of physical ones and no need to re-index results that can take hours before changes can be effective.


Beyond all the points I've already made in this blog post, let’s look at real results. For most customers with significant investments in BW, the opportunity to really leverage their mission critical data warehouse investment and unlock its full potential has always been the goal. I believe the opportunity has arrived with the marriage of BW and HANA.


The vast majority of our 50+ BW/HANA ramp-up customers have shown impressive results and fast ROI. And investing in HANA today as the database for BW has set them up to further expand the value of HANA as a foundation for real-time analytics as well as a platform of new innovative applications. What this means is HANA for BW today is the beginning of your platform for the future. The same HANA platform will enable your advanced analytics use cases, run your planning applications such as decision support/predictive analytics, text mining and search.


Also, it is a simple migration of your database and SAP has a cookbook to help you do this.



Scenario #2: I have SAP ECC and may want BW, what do I do?


The short answer: It depends.


It depends on a number of factors and considerations when looking at your overall data management, data foundation and analytics strategy going forward. For example, if you have planning needs, you can take advantage of solutions that are based on the BW foundation. You may consider BPC on HANA, S&OP on HANA as possible solutions (This could necessitate BW).


For many companies with significant investment in SAP ECC who are looking at an enterprise data warehouse approach to consolidate corporate data, provide one centralized trusted data foundation for analytics and with one version of the truth, SAP BW is the way to go. If someone tells you they can build a custom warehouse similar to BW, replicate all the "good" that comes with BW with no downside, I'd like to meet that person.


If SAP transaction systems are the primary source of reporting information for you - then it makes a lot of sense to use BW as your data warehouse to make use of its business content and layered scalable architecture. Also - there is no reason why you cannot have BW on HANA AND a stand-alone HANA system, both as part of your landscape at the same time if you have use cases that need it. Having both might enable you to determine what, if any, part of BW you might want to leverage.


That said, there are times where a company has a very specific pain, which could be better solved through a custom warehouse or SAP BusinessObjects Rapid Mart.



Scenario #3: I have SAP ECC, but someone told me SAP BusinessObjects Rapid Marts are better, what do I do?


Short answer: Again, it depends.


SAP Rapid Marts definitely do not substitute a data warehouse. (That's why they aren't called "Rapid Warehouses") But they do offer value for companies looking for a quick "plug and play" approach who need a light weight analytical datamart often to support a departmental use case. In other words, Rapid Marts should be viewed as accelerators which could be a prelude to a larger data warehousing project.


For those unfamiliar with Rapid Marts, they include pre-built data capture jobs, data mappings, a data foundation and packaged reporting content for accelerated BI implementations.  They target key areas of SAP ECC (e.g. Finance, HR, Manufacturing, etc.) and are great for departmental needs in a large enterprise or for smaller companies with scarce IT resources but with an immediate need to get fast insight into their ECC application.



Scenario #4: I have SAP ECC and someone told me they will design a data warehouse for me from scratch, what do I do?


Short answer: Designs are cheap. Deployments aren't.


If you don’t think this will be expensive, think again. As I previously stated, SAP spent the last 15 years building and perfecting data mappings from the suite to BW.


For example, let’s look at integrated lifecycle management with SAP ECC, complete transaction record integrity with ECC, pre-rationalized data models and data objectives for ECC.


Trying to accomplish all this for your SAP landscape from scratch would mean that you are essentially re-inventing the wheel and not unlocking any value available in BW - for free. In this scenario, an important consideration is SAP BW’s established design and pre-delivered packaged content (for fast time to implementation). This is why the BW option has been the route taken for the majority of our customers (13,000+) who have turned to BW as their data warehouse.


I realize that we are going to get some debate going about design vs. deployment. Like I said...design can always be done on the cheap.




So back to the original question: Does HANA replace BW? I suppose the answer is: No. But putting a statement out there like "if you haven't deployed BW you shouldn't" would be incredibly irresponsible. HANA is definitely many things (A database for BW, a high-performance analytical appliance, a platform for new applications), but matching the entire "system" known as BW point-for-point is a huge project for any company.


I have no doubt in my mind that most of you who have BW systems today would agree and some people will disagree with me (it’s guaranteed and I expect Inmon/Kimball quotes aplenty), and spurring on conversation was certainly part of the objective. I know I only covered a few scenarios and request that responses to this blog suggest more specific scenarios for a “Part 2” of this blog.


For insightful discussion on BW on HANA check out the blog posts on the Experience SAP HANA site.

*To read an update on this blog, please click here: Does SAP HANA Replace BW? (Hint: Still, No.)

Recently, I chanced upon a blog by Mark Rittman and his follow on comments to a reader. He often blogs about a variety of Oracle products and related topics. In this particular instance, while discussing potential use cases for Exalytics (http://www.rittmanmead.com/2012/06/some-thoughts-on-potential-exalytics-use-cases/#comments), he makes a couple of very keen observations when asked about what use cases Exalytics would not be good for. In his words (June 3, 2012):

“….there’s two scenarios where the Exalytics approach (as it currently is), in combination with OBIEE might not be such an obvious choice:

1) Real-Time Data Warehouses with continuously updated data, fed either by micro-batch ETL or push-type technology – in this case, the caching and pre-aggregation approach used by Exalytics won’t really be useful, as they will become stale (out of date) as soon as the DW dataset gets updated.”


A point that stands out is that there are indeed use cases where the data being analyzed needs to be truly current or real-time, and not a result of any pre-aggregation. This is an important consideration because there are indeed many cases with critical events that need addressing right away and need to be based on a more comprehensive data set; in these cases, any delay and data set limitations caused by any pre-aggregation can be damaging.


This brings to mind a certain use case that I recently discussed with a start-up company called AlertEnterprise (they participate in the SAP Start Up Focus program). They outlined a use case for what they describe as “Critical Infrastructure” customers, such as Airports, Oil & Gas Pipelines, Refineries, Chemical Plants, and Utilities. For such customers, “Insider Threat” is a matter of constant worry. Recognizing the severity of such threats, even the Department of Homeland Security (DHS) has listed Insider Threat as an Advanced Persistent Threat (APT). Insider threats pose significant danger to the systems – physical, logical, and security related – of these critical infrastructure customers. For example, at airports insiders have privileged access to airport processes and procedures, access to secured areas, and are often aware of where there might be chinks in the armor with respect to overall security. Thus, effective security at these customers requires a multi-faceted approach to address a variety of threats, both external and internal.  Recent studies and information obtained by the DHS, the FBI and other agencies, indicate that insiders are not only utilized by terrorists to gain access to sensitive information and targets, but they themselves are also often likely to be perpetrators of hostile acts. 


To effectively respond to insider threats at airports and other critical infrastructure customers requires predictive risk analytics and utilization of cutting-edge security convergence technology. Security convergence implies analyzing risks across IT security, physical security and operational systems (such as those meant for Supervisory Control and Data Acquisition (SCADA), etc.), to safeguard critical assets. The ability to do so has not been very effective in the past. Using sophisticated algorithms, AlertEnterprise seeks to solve this problem. However, the use of these algorithms alone would not solve the problem very effectively; the challenges posed by huge volumes of data, structured and unstructured (possibly from a number of disparate sources) would slow down the processing significantly. This is where SAP HANA makes a difference.


Without SAP HANA, Predictive Risk Analytics for security would take a very long time to process. In this case, this is obviously a crippling factor because, as they explained, predicting the possibility of malicious events with an eye to preventing them is pointless if the calculations would continue well past the occurrence of the incident itself!


Working with the SAP Co-Innovation Lab (COIL), AlertEnterprise has demonstrated how this process would leverage some of the key differentiating aspects of SAP HANA, namely, its tremendous computing power and the ability to rationalize large data sets from diverse information sources; and, just as importantly, its ability to deliver on these fronts without having to massage or pre-aggregate any of the data.  The solution in this case demands the ability to process data from a variety of identity databases. For example, in the case of airport security the Transportation Systems Clearinghouse, No-Fly Lists and HR systems for airports would be sources of data. This is important to note because insider threats come in many shapes and forms at such critical infrastructure customers, and each one of these could potentially represent a unique data source. Thus, the raw processing power of SAP HANA, combined with the ability to deliver without any pre-aggregation or pre-selection of data provides true real-time results – that does the job! Today, it appears that SAP HANA uniquely provides this powerful combination of capabilities because many of the other offerings on the market that claim to provide blazing fast analytics do require some form of data pre-selection, pre-aggregation, etc.


As I see it, this strength of SAP HANA is but one facet of the comprehensive thinking behind its design, a design that represents nothing short of a revolution in the world of computing. Keep an eye on its rapid evolution and ongoing enhancement of capabilities, for that will help you identify newer ways of solving old problems and effective solutions to new ones!


P.S. You can follow my other blog posts at: Café Innovation Blog on the SAP Community Network (SCN)

SAP NetWeaver Business Warehouse (BW) has become the standard platform for Enterprise Data Warehousing for companies of any size with more than 16.000 productive environments and constantly growing at 200+ new installations each month.


With BW on HANA, SAP provides our customers a performance boost for their exiting environment along with new capabilities like real-time access to SAP HANA Analytic Views and accelerated Integrated Planning.


This performance gains boost both efficiency and time-to-decision across many lines-of-business and IT, which creates tangible value in various business processes. Customers are already  reporting process improvements in areas like:

  • Financial reports based on various sources of data can be compiled in a more efficiency manner
  • Sales representatives spend less time on generating reports and more on actual sales activities
  • Information requested by external auditors can be delivered more efficiently and quickly


Also the IT departments largely improve operational excellence by:

  • IT needs to spend less time genera-ting and compiling management reports
  • IT is relieved from having to many data aggregation activities and is capable of running data-load-processes more efficiently


A business benefit framework should contain both the improvement boosts for the reporting business processes and for IT operations. The graphic below shows an example how BW on HANA can contribute to improvements of processes, BI and IT infrastructure:


In the next blog, we will show you how to setup a proper benefit model, explain the baseline figures and estimated improvements based on customer installations an internal measurements.

If you are looking for ideas to apply SAP HANA within your organization then the HANA use case repository hosted on experiencesaphana is the perfect place for you to explore and be inspired! I launched the HANA use case repository at Sapphire Orlando with 50+ use cases by featuring it at the discussion table “Building the Business Case for SAP HANA” within the Database & Technology campus. If you did not get a chance to step by the booth, you still have a chance to listen to Steve Thipodeau from SAP Value Engineering explain how the repository was used to engage with customers related to SAP HANA. 

Check out the following new and updated use cases

The true value of any resource such as this repository comes from how regularly content is added or updated. As such, Sapphire offered exciting new content that I could harvest as more than 30 HANA customers shared their experiences and usage of SAP HANA.  So  even if you have already browsed the repository here are three examples of new and updated content that is waiting for you to explore and why you should re-visit often in the future:

  1. The Genome Analysis use case articulates how by using SAP HANA it becomes possible for organizations such as MKI to deliver advanced medical treatment to many more patients
  2. The Real-time offer management for on-line gaming use case is a great example of how HANA can enable gaming organizations to capitalize on business opportunities not possible before and highlights how industries such as online gaming become very much relevant for SAP and its eco-system. Listen how Bigpoint has used SAP HANA to monetize online gaming.
  3. For the Real-time financial planning use case you explore the great new resources that have been added. Watch Colgate talk about the opportunity for BPC on HANA within their organization, how ConAgra expects to reduce the integrated business process cycle time from weeks to days and how Lonmin can now run more planning scenarios simultaneously thereby improving their decision processes.


Explore use cases and share your insights

The  graphic below shows how you can easily find and browse use cases for your Industry or Line of Business. Also, if you are aware of a good HANA use case, which is not yet described in the repository, then I invite you to share your insights by clicking on the Share your use case button. 

use case repository image.png

All it takes one click to get started http://spr.ly/HANAUseCases.

Since the last time was such a huge success, we are rolling it out again: SAP, Palo Alto, is hosting another SAP Startup Forum – the same and yet different – this time we are focusing on Mobile on HANA and it is happening on Wednesday, June 14, 2012. Even though we’ve been around for 40 years, SAP still has the passion and heart of a startup, and startup forums are a great example of that mindset. Our objective is to tap into the Silicon Valley ethos and link some exciting mobile startups with key SAP technologies and decision makers.


It’s only been 12 months since the introduction of SAP HANA to the database world. The revolutionary new database architecture from SAP, is an in-memory break-through and has captured the imagination and innovative efforts of large enterprises globally. But the reach of HANA has impacted the collective database community all the way out to innovative startups. At SAPPHIRE NOW 2012, 10 startup companies showcased their solutions running on HANA and the SAP Startup Forum on June 14 will explore startups HANA potential on mobile.


Startup forums are a day of collaboration and learning and the June 14th event will focus on mobile trends and technologies as they relate to big data and SAP HANA. Selected startup companies will have the opportunity to present to a group of attendees from local media, SAP and SAP Ventures as well as exhibit to SAP employees in a tradeshow setting.


Invited startups will contribute to the mobile on big data discussion, engage and learn from SAP technical resources, hear from and meet with key SAP leaders, explore customer and funding opportunities, and network and win ‘bragging rights’ over the course of the day.


If you want to follow along virtually follow #SAP #Startups for all tweets related to the event or @SAPInMemory for all HANA related and @SAPMobile for all mobile related news. Feel free to ask questions by using #SAP #Startups.


If you are interested in participating in a future SAP Startup Forum or want more information on the program in general you should click the “Start Here” button on http://spr.ly/SAPStartups .


See what the last startups experienced with SAP HANA and how it played a critical role in solving complex business scenarios.


Row <it>and</it> Column

Posted by David Dobrin Jun 5, 2012

Before I Begin, An Apology:

This blog post tries to correct an error I made in a previous Experience SAP Hana post, an error that was remarked upon frequently. I said in the piece that HANA could be "either column-oriented or row-oriented;" and that there is "literally a switch" that you throw. Right about the row <it>and</it> column-oriented. Wrong about the switch. "David, you know there's no magic wand that you can just wave," one of the original designers told me.  "It's just not that simple."


In this piece, then, I'll try to bite the bullet and explain how it "really" works, that is, how a single database, containing tables with rows and columns, be both row <it>and</it> column-oriented? Please note.  I'll do a lot less hand-waving than I was doing when I said, "there's a switch."  But hand-waving will not be entirely eliminated.  And there may come a point where you might long for a little oversimplification.


Column AND Row

Let me start with stuff that we all know.  HANA is a database.  In the database, there are tables. The tables are essentially blank spaces where you can load collections of fields called records.  In a row-oriented database, each row in the table contains a record, and each field in the record occupies space, even if there is no actual data there.


Here's an example.  Imagine a table of home addresses, which contains places to put the last name, street, city, and telephone number for a number of people. (Each of these "places" is called a field.) In a row-oriented table, the filled-in table would contain many lines (or "rows"), one for each address. Each line would contain all the values you had for that person, but if, for instance, you didn't have a telephone number, the space for a telephone number would be left blank.


In a column-oriented table, each field occupies a separate column.  When you load an address in, you don't put everything on the same row.  You just add any actual value you have to the end of the column that it belongs to.  So you add the last name to the "Last Name" column, the city to the "City" column, etc.  And if you don't have a value, you don't add it. 


So how does the column-oriented table know that data which is now in different rows actually "belongs" to certain other values that you loaded in? (The technical term is "record.") Well, essentially [handwaving alert!!] each value is tagged, and there's another table where you use the tags to look up which record the value belongs to. 


You can see that a row-oriented table is much easier to understand conceptually (it's like rows in a spreadsheet), but a column-oriented table is much more compressed and, as it turns out, much faster if you want to get data back out again and work on it. 


In a really interesting paper on HANA (Sikka, et al., "Efficient Transaction Processing in SAP HANA Database – The End of a Column Store Myth,"), Sikka characterizes row-oriented tables as "write-optimized," and column-oriented tables as "read-optimized," and this is certainly a fair description of the two kinds of tables' essential strengths. Use rows for transactions, where you're trying to enter data.  Use columns for analytics, where you're trying to look at data.


It would seem that this choice (row <it>or</it> column) is just built into the nature of the world.  It's either in rows or it's in columns.  And perhaps for that reason, most databases make a choice about how they're going to operate.  Either all the tables in the database are going to be row-oriented, or all the tables in the database are going to be column-oriented.


But not HANA.  In HANA, a table can be row-oriented or column-oriented; the programmer (or data designer) simply decides which it is to be. 


But that's not all there is to it.  Because, for a programmer, inside HANA, the choice of row <it>or</it> column is not the same as it would be elsewhere.  In HANA, if you decide that a table is to be column-oriented, you don't necessarily give up on write-optimization (what row tables are good at).  HANA has set up an internal system for column-oriented tables that essentially [hand-waving alert] allows you to enter data into a buffer that is row-oriented, then, in a series of steps, push the data into the column-oriented table, without impairing the performance of the column-oriented table.


So, in two quite different senses, HANA is both row-oriented <it>and</it> column-oriented.  A.  It allows you to create either row-oriented tables or column-oriented tables.  B.  If you create column-oriented tables, the column-oriented table has a row-oriented buffer for data entry which is write-optimized, so effectively [hand-waving alert] the column table behave as if it were a row table.


Write-Optimizing the Column

It's worthwhile to understand how this is done, because it gives you some insight into how HANA is using in-memory. 


The following diagram is drawn from the aforementioned paper. 



Figure 4.  Overview of the unified table concept

Source: "Efficient Transaction Processing in SAP HANA Database - The End of a Column Store Myth" Vishal Sikka, et al., pp. 731-741.

The columnar table in HANA can act like a row store for one simple reason.  It is a row store.  As you can see from the diagram above, there is a buffer store (L1) which is row-oriented.  Data coming in goes into L1. 


Every so often, an L1 buffer is effectively closed, and the data in it is inserted into a column-oriented buffer (L2).  This involves the typical column-oriented things, where the values are added to their respective columns and the fact that they belong to a particular record is recorded in another table.


Every so often, the L2 buffer is closed and the data is pushed into the primary columnar store. 


To get the data into the columnar table without hurting performance, it first puts the data into a "write-optimized" [row] store (L1). It then takes that data and restructures it, pushing it into a second store (L2), which is column-oriented. Finally, it pushes the data in L2 into the main column-oriented table which was the intended destination all along. 


This process would be weirdly byzantine and slow if it weren't done in-memory (and in parallel).  But since L1 and L2 and the final column store are in memory and all the processing needed to make this work is done in memory, it can all be done in the background. And because L1 and L2 are relatively small, the data in them can be made available to other database operations. 


So, for instance, if you run a query on the columnar database while some data is being added, the query just checks L1 and L2, as well as the primary store Since L1 and L2 are always relatively small and since the query is done in-memory, the performance cost of doing this is small.


When Do You Use Row, When Column: One Man's Opinion

You can see, now, why I chose to do some hand-waving when I was explaining the design idea behind HANA.  L1, L2, write-optimized, blah, blah,  Important to know if you want to know what it does.  But hard to grasp if you're just trying to get the idea. 


Besides, a little learning is a dangerous thing.  When you know things, you start asking questions.  Why provide that much flexibility?  What is it good for?  How will people use it?  What does it enable?  If we were working at the hand-waving level, you just wouldn't be asking questions like that. 


But now that I've given you the detail, I have some responsibility for explaining what this row-and-column stuff is doing.  I myself can't answer from my own experience.  So I did the next best thing.  I asked people who have built products in HANA


I started with Adam Thier, whose job title at SAP used to include the words "database" and "architect." 


Why or when do you use row-oriented tables and when do you use columnnar tables?


"I am a financial software lifer [he worked on Essbase at Hyperion and is now working on EPM at SAP]  in a long term platonic relationship with boring algorithms like "depreciation" and "allocations"; and I've ground my teeth to nubs trying to make basic algebra fly in a relational database.  That's why I jumped ship to Columnar back in the 1990s and never looked back.    I typically always start there with the exception of long text or Blobs (for things like descriptions).


"However, even in the world of HANA, that's not a constant because of the wonders of text analysis."


So, if you have to do calculations on the data, you use columns, and if you have to do other things, like store text, then you use rows, except when you don't.


That seems clear enough.


When Do You Use Column, When Row: Another Man's Opinion

Not content with that, I also asked Vijay Vijayasankar, who has done a fair amount of HANA development and has served as an InnoJam judge with me.  He's the sort of guy who would probably prefer to post on this himself, so I'll just pick and choose from his answer.


"HANA can have column or row stores; there is no technical limitation.  My simple rule of thumb is 'use a column table until proven wrong.' 


"All data is going to be stored in more than one table, so you also need to think about how you will join the tables.  The worst performing apps are the ones that try to join columnar and row stores.  The reason for the drop in performance is that the result of such a join needs to be temporarily materialized, which kind of defeats the purpose of HANA.'


"There are other constraints.  You cannot build a column view if the underlying table is a row table.  A row store table is loaded into memory the moment the HANA engine starts, whereas a column gets loaded into memory when a query needs it.


So again, you use columns, except when you use rows, except for when columns are better.


So Does It Matter?

I went back to Adam with a bit of a puzzled look on my face, and said, "?"


"The beauty is that with HANA it doesn't matter. I can start columnar, and if I stumble I can spend a few hours and reconfigure certain tables over to rows, with a smile on my face the whole time." 


So, in effect, not all that much has changed in HANA?  You did columns in the past, and now you're still doing columns?


"No. I'm doing columns in a different, more transactional way.  Remember, I -- and a whole bunch of people much smarter than me -- who go back to the original MOLAP engines (essentially column stores) never cracked the code on a hybrid database -- which meant we never cracked the code on how to do an insert in a MOLAP structure without then taking it off-line to re-index. This isn't a big deal -- it's a huge deal."


Simple and Fruitful

In my earlier blog, I said there were too tests to apply to the design of HANA.  Is it simplifying?  And is it fruitful? 


This post might seem to suggest that there's a problem with the "simplification" criterion.  But I think that's a deep misreading of the situation.


To begin with, it's clear that even in the simple case, where columnar tables are being created because what you want is ultrafast analysis, the underlying design that I described in the first blog post makes a material difference. 


* In this design, for instance, the columnar tables can be updated continuously, whereas they can't be in most analytic databases.  (You need to take time out to reindex.) 


* In this design, too, the creation of views--like the multi-dimensional view you see in most analytic databases--is logically separate from the table structure, so you can do lots of different views, including views that do algebra, without rebuilding the tables, and you can change them as much as you want.


* In this design, as well, you can embed things like statistical functions right in the database, whereas in most analytic databases, you have to pull the data out to do a lot of statistical analysis.


And the more complex cases?  I'm not sure we know the answer yet.  What we do know is that at least two sharp database guys are pretty much saying, "It depends," when you ask then how to use the "row and column" structure.  That could be because the structure isn't all that useful, but it could also be that the flexibility that Adam describes gives you a lot of options.


At this point, should one have an absolutely clear idea of how people should use this flexibility?


I don't think so.  Certainly, you saw a similar fluidity in original electric car example.  The basic design idea was there in the battery-powered ignition system for the 1912 Cadillac. But at the time, no one was quite sure about where it was going to lead.  The original designers were mostly just trying to get rid of one limitation in automotive design, the fact that you had to turn a crank to start the car. (According to legend, the inventor came up with the idea after a wayward crank had broken the jaw of a friend of his, who later died of gangrene.)  They certainly saw that there was more to the idea than that.  But it would take at least ten years of experimentation before the standard modern automotive electrical system would emerge full-blown. 


About the Author

I run a small analyst firm in Cambridge, Massachusetts that does strategy consulting in most areas of enterprise applications.  I am not a database expert, but for the past year, I have been doing a lot of work with SAP related to HANA, so I'm reasonably familiar with it.


SAP is a customer, but they did not commission this piece, and they did not edit or offer to edit it in any way.

Would you like to identify hidden revenue opportunities within your customer base, devise offers to retain your high-value customers, partners and employees, or have your call center agents delight customers with the best next-step recommendations?


If the answer to any of these is yes then read on…


As you probably already know, we recently announced general availability of SAP HANA SPS4. Launched during SAPPHIRE NOW, this represents an important update for SAP HANA not in the least because it includes significant new predictive analytics and data mining capabilities. Several new algorithms are being added to the predictive analysis library (PAL) and fully exploit SAP HANA’s in-database processing power. SPS4 also introduces integration with open-source R to add server-side processing for the wealth of algorithms available in this extremely popular statistical and data mining toolset.


SAP BusinessObjects Predictive Analysis, our new predictive modeling and visualization application, integrates directly with the predictive analytics in SAP HANA to allow you to fully unleash the potential of big data. Or you can simply embed PAL or R algorithms into your own SAP HANA applications - in fact that’s exactly what we’re doing ourselves at SAP with solutions like SAP Smart Meter Analytics already incorporating PAL algorithms.


So what’s the reaction been to all this new stuff? Well here are some quotes from videos made by some of our early adopters:


“The HANA platform at Cisco has been used to deliver near real-time insights to our execs, and the integration with R will allow us to combine the predictive algorithms in R with this near-real-time data from HANA. The net impact is that we will be able to take the capability which takes weeks and months to put together, and deliver just-in-time as the business is changing.” Piyush Bhargava, Distinguished Engineer IT, Cisco Systems


“Our solution is to incorporate SAP HANA along with Hadoop and R to create a single real-time big data platform. Data mining will be handled by R and assisted by HANA. Data pre-processing prior to data analysis and high-speed storage will be managed by Hadoop. With this we have found a way to shorten the genome analysis time from several days down to only 20 minutes.” Yukihisa Kato, CTO and Director of MITSUI KNOWLEDGE INDUSTRY


You can watch the full videos here: Cisco, MKI.


If you’re excited to learn more about predictive analytics and how it can help you extend the reach of analytics beyond simply knowing “what happened” to proactively predicting and influencing future outcomes please visit the HANA Technology area on ExperienceSAPHANA.com. We’ve uploaded some great content to help you learn, hear and see this for yourself including a comprehensive overview and demo.


Stay tuned for further updates and we’d love to hear your views and experiences on the subject of predictive analytics.

News Archive

SAP HANA News Archive

Filter Blog

By author:
By date:
By tag:

Contact Us

SAP Experts are here to help.


Our website has a rich support section designed to help you get the highest quality answers quickly and easily. Please take a look at the options below to get you answers from the SAP experts.

If you have technical questions:

If you're looking for training materials or have training questions:

If you have certification questions: