Hana Can Do More


The fear is that Hana will only be used as a quick but mostly unimaginative SQL database. Every ERP system needs a database: for R/3 and ECC 6.0 it was Oracle, IBM DB2, MaxDB, Microsoft SQL Server, among others. For S/4, only Hana is available to SAP customers.
SAP likes to emphasize that Hana is not a database, but a database platform—which is increasingly true. The platform includes not only a powerful SQL database with extended language support, but also a graph and vector engine.
An SAP Knowledge Graph is currently being created from the "graph theory" of the Hana database platform, and the Vector Engine can be used to evaluate LLMs (large language models), but graphs are not limited to only that. Graph theory is an important branch of computer science, and many programmers are familiar with its methods in the form of Petri nets.

Graphs can often be used to represent relationships and connections more quickly, easily, and visually than SQL tables. This relationship pattern is also very easy to compute with because graphs can be converted to matrices and there are numerous libraries for matrix computation.
Representing multidimensional data in vectors may seem complex at first, but the work is worth it. When products with many properties are represented as multidimensional vectors, the Euclidean distance, and thus the similarity and clustering of product groups, can be determined almost as a head calculation. What at first appears to be an non-homogeneous mountain of data can end up as a tidy structure of a few clusters.
Hana is thus capable of managing more than just SQL tables. Unfortunately, there is no scientific research on the extent to which the Hana options are being used and where the Hana roadmap might lead SAP's customers.
SAP itself wants to build a knowledge graph for the Business Data Cloud (BDC) using its own graph engine. The goal is to create a semantic layer that could facilitate the orchestration of different ERP data silos.
A result based on the Hana database platform consisting of SQL and a graph and vector engine could become a digital twin of an ERP system. After the semantic layer of the Business Data Cloud, the Digital ERP Twin would be a powerful design tool for modeling and orchestrating a Next Gen ERP. This makes Hana, BDC, and BTP (SAP Business Technology Platform) seem more relevant and meaningful than S/4, which is perhaps only an intermediate step towards a true AI-based ERP. However, a future composable ERP will have Hana as its final database platform, proving that Hana can do more than SQL.