SAP is doing a nice job spreading the phrase “big data analytics” through the market with HANA. However, the “concept” HANA is more than an in-memory database. HANA could help you to speed database processing to create cubes with tens of dimensions and thousands of millions of records from hours to seconds, and take the results and process them again and again and again… Just try to do that with any conventional DB with so many records, would you be able to achieve the task?, the answer simple, you would not.
HANA doesn’t help you to define the cube and its post-processing logic. You need skilled people to do that. And these people need otherones to tell them how to design the cube; those ones are data scientists.
Data scientists require understanding of what kind of resultant data you desperately need to obtain better profit, or more revenue, or to lower costs. Then, probably you are going to have to provide access to your main databases, documents and strategic information, the rest probably could be gotten from internet (weather, social network, press releases, blogs, competition).
Sometimes you need an intermediary, a consultant specialized in your market industry, between the data scientist genius and your decision makers or influencers to smooth the interaction. Sometimes smart guys need to help you to communicate with the rest of the world. And this consultant can help you to find the right dashboard that you are expecting to have on your hand and see how to improve gross margins, lower the costs, speed business cycles, etc.
Data scientists are going to take all this information and define a framework to run on this data with different elements such as algorithms, probability models, pattern recognition, decision trees and other things hard to be followed; and call the HANA guy to build the required cube to get the process done. Now, you need a data service guy to install an ETL or whatever it is, to keep HANA fed with updated data.
When this part of the setup process is ok, then you have to start preparing the delivery the results in a human friendly way. This, for example, could be done through your Tablet. Then you need to involve an additional team to build the visualization.
Also you need a deep understanding how to monitor and manage HANA in a flexible and scale-out infrastructure in order to stay elastic and react in time to data growth needs. It’s not just to buy the stack that some vendor sell as a HANA certified one. Do not bother to install this in a server or appliance, this will be one of your most important assets, you must install it in a high available configuration.
HANA needs maintenance, and that means people who need to take care of memory availability and move tables among nodes to balance performance.
SAP provides resources to speed the construction for some use cases. However, every customer has different needs, and then you usually have to add specific visualization formats, or more data to the analysis that requires a different treatment.
HANA also helps not just to analyze, also supports under some conditions predictive, this case usually requires in-memory post-processing with tens of interactions of the same data set, every touch in your tablet produce this bunch of interactions of thousands of millions of records with tens of dimensions each one. I bet you will be excited to get results in seconds for every touch all times. I also dare you to do that in a conventional database.
The point of this note is to show that you are not buying a database acceleration product; you are buying a full set of skilled professionals and resources working as a team to provide tools and real-time information to get much better productivity, profitability or change the course of your business.
Usually you don’t need to accelerate your primary database, everything depends on what you really need. Sometimes you discover amazing things during the construction process with some earlier results, and I bet your appetite will come up for more.
Anyway, it’s truly important that you got involved throughout the setup process, don’t miss the opportunity to find your next core business or your most important competitive advantage. On the other hand, despite the initial implementation might be hard, it’s easier to add more capabilities in the future time with so much less investment to go for the next indicators that can provide you the differentiation that you were looking for and I am pretty sure you will do it because any competitive advantage is temporary.