“Game over for SAP BW – Game on for SAP HANA!” This is what many of us thought when SAP launched its in-memory database in 2010, and its marketing machine was droning on and on with promises of reporting and data updates on steroids. Confusion reigned almost supreme: Marketing told us that HANA was more than just a database and HANA Studio seriously looked like an alternative to SAP BW!
As a BW consultant, I was thinking: ”So ein Ding muss ich auch beherrschen!”, which is why I took the ”TZHANA: SAP HANA In-Memory Appliance Software” course in 2011. At the course, I met a guy called Richard, an employee of an English food manufacturer that had gone “all-in” on HANA – they were about to get under way with migrating the BW solution to HANA, which was also the reason why Richard was attending the course. As the course progressed, it gradually became clear that HANA was not a substitute for BW in the form of a centralised Enterprise Data Warehouse as it, for example, lacked management of authorisations and time-dependent data. Basically, if you had chosen to send BW packing like Richard’s employer, then you were forced to build your own EDW application. Even though the idea of building everything from the ground up was exciting, I did not envy him the task he was facing…
Out into the open!
It took just under a year for SAP to lay the cards on the table and explain that:
EDW = a database + X
The database = SAP HANA and X = SAP BW in the equation.
Well then! In other words, from a BW perspective, HANA is just a database that has a number of wild tricks up its sleeve. “BW on HANA” started to shape up, and it is now, at the beginning of 2015, that SAP is about to give its BW clients, who have grown in number, a mature offer in the form of “BW 7.4 on HANA“.
The vast majority of BW implementations function as central EDWs where users access data using one or more of SAPs BI Frontends. Data update, reporting and budgeting run times continue to pose a challenge. SAP HANA solve this. Well, it is not strictly necessary for your P&L report or your budget to be updated in a split second – after all, it is not operation-critical information we are talking about here – but it is extremely frustrating as a user to have to wait 30, 60 or more seconds for the screen to update. We have all grown used to Google where everything happens in the blink of an eye. And this is what your BW solution should be capable of, as well. There is no way around that, and even if it is still relatively expensive to throw a SAP HANA database under a BW system, this is going to happen.
How do you migrate your BW to BW on HANA?
You must therefore decide how to go about your migration, and here you have two options:
- A “technical migration” where the “only” thing that is replaced is the database and where your BW objects are converted to being “HANA-optimized”.
- A complete re-implementation of the existing solutions in a BW and HANA environment.
The first option sounds immensely alluring as it generates no interruptions or inconveniences for the users and can be realised relatively quickly. A disadvantage here is that you will need to transfer your budding BW solutions (they exist, you see, in your system, too!) and make corrections to, e.g. the reference routines in order to be able to take full advantage of the HANA database. Then you will need to fill your dearly-bought database with a layer of unnecessary data in the form of BW cubes. In many cases, you will want to get rid of the cubes and report directly to the data layer underneath the cube (however, not for the inventory, stock and planning cubes, i.e. the budget and forecast solutions). I would imagine that many will go for this solution, simply because it guarantees improved performance in many areas in no time, thereby generating quick and, in particular, stable profits. And then you just have to accept that you will need to tweak a couple of things here and there in individual solutions to derive reasonable benefit from the database migration or to remove unnecessary layers of data.
The other option requires the ambition to take advantage of the golden opportunity to have full knowledge of how data are transferred from the source systems, how they are processed and how they are handed over to meet the company’s needs. The disadvantage is that this is highly likely to require a greater effort and it will also take far longer before you can go on the air. Add here the company’s efforts to regression-test all solutions and the need to have the old and the new system running in parallel for a long period of time.
I exchanged a couple of emails with Richard from the HANA course about a year ago and found out that they had ended up scrapping the plans for replacing BW with HANA and went instead for a full re-implementation of BW on HANA. This was the ultimate result of their large-scale upgrade project. In other words, they went for option no. 2. A probable explanation for this is that the whole organisation had already become reconciled to having to go through a re-implementation exercise, so there was both budget and backup for this. Teething troubles there were plenty of, but the result was terrific: data update, reporting and budgeting run all like greased lightning. But what is most important of all: the internal BW department is fully aware of why the different solutions are screwed together the way they are, and this knowledge is scarcely economically quantifiable!
Try it out!
If you would like to try out how a BW system works on a HANA platform, SAP offers a 30-day free trial. What we offer more exactly is a “BW on HANA cloud”. You are required to have an Amazon Web Service account that costs mere pennies to run – provided that you remember to deactivate the application when you will not use it. I have been toying with this offer myself over Christmas in order to see how budgeting works on a HANA platform – and the result is splendid! Enjoy!