Ingeniously off the mark!

There is an unwritten data law at SAP: No duplicate data storage! At first glance, this sounds very reasonable; in practical IT life, less so. For Hasso Plattner, it is an IT dogma.

Looking at the successful history of SAP, it is obvious: The central database server in the three-tier client/server model is the essential architectural component.

With Hana, SAP has now even gone one step further: Hana is not just an SQL database, but as a platform the infrastructure for ERP, CRM, logistics, SCM, etc. Hana goes beyond the pure function of a database - but this does not make the demand for central data storage any easier.

Plattner's dogma of centralized data storage in the first approach is also independent of the technology, whether the database is disk-based or in-memory, and also independent of the infrastructure, whether on-premises or cloud computing.

But SAP also had to recognize that any rule can only be as good as real life allows it to be. In a heterogeneous IT landscape, the dogma is very difficult to maintain.

Compromises become necessary. Ultimately, "centralized data management" mutated into an esoteric concept that does not hold up in practice.

Because SAP also understood the challenges of different data silos, they developed the concept of SAP's data hub - again, on paper and in theory, a stroke of genius: Since the singularity of data storage cannot be maintained, SAP is trying a "linked" system of references.

Furthermore, the data itself is not moved, but remains unique in its silos. If an app needs the data from another app, there is no "copy and paste", but a reference and the asking app now "feels" in possession of the data! Ingeniously off the mark!

For this reference system to work, the requested app must handle all the "data preparation" for the asking app - after all, the data itself is not supposed to be moved.

The requesting app will have no problem with this, but the requested app might show symptoms of stress: If the requested app receives many data requests, the workload increases dramatically and the app's server only works on behalf of numerous requesting apps.

Is this the last word in wisdom? Interestingly, SAP itself has better concepts from the open source area, but they are often only mentioned in passing.

We encouraged our base people to get involved with open source data concepts very early on. The success proves us right, as does the existence of an SAP Hana Spark Connector. Thus, my base people not only regularly go to SAP TechEd, but also to the relevant open source events such as Cloud Foundry Summit, etc.

TechEd: My base people told me that there is a new API layer called SAP Graph, based on SAP Graph, but not SAP Graph! Brilliantly off the mark!

In my very long ago computer science studies I had a lecture on graph theory - very interesting, but at that time also very theoretical. My astonishment was therefore all the greater when I discovered a graph engine in Hana a few years ago.

In addition to relational tables, Hana also masters the function of a graph database. Now SAP is apparently building a semantic data model on top of this graph engine, which is intended to consolidate access to SAP data as an API - similar to SAP Data Hub.

This API layer is called SAP Graph to definitely irritate all SAP Hana fans and basic people as well as app developers - brilliantly off the mark!

Because I've chosen "Ingeniously off the mark!" as my motto this time, here's a quote from former CEO Bill McDermott:

"I have at least thirty more years to work. I very much enjoy being a CEO. When I'm eighty and have enough, I can always think about getting involved in politics and society."

He said in an interview with the FAZ on April 10 of this year.