The Hana Database Alternative

After APO came Hana

SAP's invention of APO (Advanced Planner and Optimizer) was a revolution in theory. The idea of automating and optimizing MRP runs was the right one, but the volume of data from production overwhelmed the APO algorithms. As has often been the case before and since, Professor Hasso Plattner had an idea that saved the day: a small, sophisticated database hidden in the main memory of the APO server, which accelerated the system and delivered the planning figures and their optimization to the SAP users' screens in a timely manner. APO was saved!

At the time, SAP did not make much of a fuss about this step in technical support. The goal was to make APO a success and consolidate SAP's ERP dominance beyond finance and controlling. APO's in-memory computing database was therefore seen as an auxiliary function not worth mentioning.

Potsdam reinvents ERP

At the Hasso Plattner Institute at the University of Potsdam, Professor Plattner and his students are thinking about what a future ERP might look like. A central idea was a new response time in real time, i.e. an ERP system that does not keep the user waiting—see the APO disaster.

Once again, Hasso Plattner leaned on the trick of in-memory computing for a database. This time, however, the database was not meant to be hidden in the depths of the ERP system, but rather was the star of the SAP community as an independent system! The new database had to be universally available across all SAP modules and systems, and it needed to be able to provide answers in real time. An in-memory SQL database was developed at the Hasso-Plattner-Institute.

Hana-only versus AnyDB

It was a brilliant move by SAP to offer Hana exclusively as a database for the ECC successor S/4 with the open source operating system Linux. With SAP R/3 and Business Suite 7 (ERP/ECC 6.0), not only did SAP experience business success, but so did the operating system and database vendors. Oracle and IBM, in particular, enjoyed high database sales without any significant sales effort.

More than 20 years ago, Hasso Plattner said at a CeBIT press conference that he would focus on the entire value chain —from the desktop to the server. After many years, we can now see the result: from Fiori to Linux and Hana to S/4 and the SAP BTP IT platform, the claim to power is seamless. Or is it?

Good sales, poor performance

In the current transition phase from ECC with AnyDB to S/4 and Hana-only, SAP is making an extremely good profit because many customers must continue running their Microsoft, IBM, and Oracle databases for years to come, while at the same time paying for pay for the corresponding licenses for Hana. Many SAP customers are currently paying twice for their databases.

Of course, Hana cannot deliver on the promise of a real-time ERP system. Those who can afford it now have all their data in the main memory of the S/4 server, which accelerates analytical queries sensationally. Any data warehouse solution benefits from Hana in an unprecedented way. SAP users rejoice! But just as many ERP functions are independent of a theoretical database speed. Other factors determine the response time, so the extremely expensive data storage with Hana in the S/4 main memory does not guarantee any benefits. SAP customers are paying for expensive Hana licenses without any discernible benefit.

Democratization through BTP

Hana has its best moments in the analytical area, so a hybrid system could be the optimal solution for SAP customers: Hana for the data warehouse and a low-cost, traditional database for the operational and serial ERP functions. Isn’t that an ideal solution?

This hybrid data management does not exist yet, but the SAP Business Technology Platform with the functionality of SAP Datasphere and open source add-ons such as Kafka-Apache could pave the way. A lot is happening in data management: Data Fabric, Data Lake, and the aforementioned Kafka-Apache offer SAP customers on the hyperscaler platforms relevant alternatives to Hana-only. Even ERP extensions and modifications using CAP, RAP and Steampunk are not solely dependent on Hana. If you want to escape the Hana lock-in, you should take a look at BTP, Datasphere, Steampunk and open source technologies.