What are Graph Databases?

Graph databases store data as nodes (entities) and edges (relationships) instead of in tables with rows and columns. This allows highly connected data and complex relationships to be represented and queried efficiently. Well-known examples include Neo4j and Amazon Neptune.

Why does this matter?

Business processes consist of diverse, interconnected objects such as orders, invoices, and customers. Graph databases represent this reality more naturally than relational models and are the technological foundation for modern approaches such as Event Knowledge Graphs and Causal Process Mining.

What does this look like in practice?

The question "which suppliers are attached to orders whose invoice has been open for more than 30 days?" needs several joins across order, invoice and supplier tables in a relational database. In a graph database it is a walk along existing edges — the relationship is stored, not computed at query time. The gap widens with every additional hop.

Not to be confused with

Not the same as a knowledge graph, which is a data model; the graph database is the technology that stores and queries it. Nor the same as a relational database with foreign keys: there, relationships are implicit in key values, while in a graph they are explicit objects with properties of their own.