Are you asking right questions about In-memory computing (e.g. SAP Hana)?

You have invested heavily in your technology solutions that are running fine. While your users complain about performance and usability of your apps from time to time, you are not sure if you want to invest in a big change. One day, you are introduced to In-memory computing by a sales exec who claims that In-memory computing can deliver tremendous performance improvement (seconds as compared to hours) to your processes. You are impressed but not sure if the increased speed of transactions is worth the cost and efforts of moving to In-memory computing. And the discussion ends here. However, the discussion should have started there by asking next set of questions. What are these questions?

While performance improvement is the fundamental benefit of bringing In-memory computing, what you would be able to do with that improvement holds the key to business value and transformational change. Can In-memory computing go beyond the promise of a faster database and deliver tangible value and competitive advantage? Let us explore the answer to this question.

For a long time, we had to deal with multiple databases. Transaction processing was run on OLTP database optimized for writing function, while analytical processing was run on OLAP database optimized for reading and analytics functions. This architecture was essential because having a single database would have slowed down transaction processing. But now with the drastically upgraded performance of In-memory database, a single database can handle both transaction and analytical processes. What is the benefit of a single database? The single database will make real-time reporting and data processing possible. As a result, you do not have to wait for days to get the data moved from OLTP to OLAP database to check your KPIs and reports, and hence those vital analytical insights can be made available in real time. Imagine what you can achieve with readily available real-time processes and analytics.

The possibilities of optimizations do not end here. Let us look at another limitation of software applications built on a traditional database. To optimize performance, these applications were segregated into multiple modules and tables to manage different functions. For examples, SAP ECC kept GL, Fixed Asset and Controlling data in silos. While this data model optimized performance of each module, it increased the number of steps and manual processing required to reconcile these modules. With SAP Hana and In-memory computing power, it is conceivable to have all the modules report from a single table (e.g. SAP universal journal for financial transactions). Manually reconciling all the modules at month end closing can finally be the thing of past. Just think what you can achieve with this massive simplification of processes and reduced steps. Moreover, the merging of tables has added benefit of reducing data footprint because you only need one record in one table as compared to four records in four different tables which reduces storage requirements. 

Do the possibilities end here? Once you start to think what all can be simplified when performance optimization is not the main consideration of your software architecture, you will start to see many opportunities. You can continue to simplify your data model and deliver simpler processes and analytics. For instance, you no longer have to maintain total and aggregate tables. Why? Remember these aggregate tables were only required so you can execute reporting quickly and do not have to rely on slow computations at the run time. With easy access to data in RAM, you can build aggregates on the fly. This enhancement not only offers real-time access to summarized views but also provides much more flexibility because you are not just restricted to pre-defined fields in total tables. You can now create ad-hoc fields when you want and modify them when you want. 

As you can see, In-memory computing is the inflection point that will enable you to implement next set of software innovations. Today, this paradigm shift in computing is at the early stage of innovations and early adopters will be rewarded by developing the competitive advantage. Of course, this transformation will only be a reality if early adopters can get this change right and bring value from this technology by asking right questions and implementing right innovations.