5 Data Architecture
5.1 Overview of data sources
The BPA Lab integrates data from several sources — from master data in a relational database, via process state maintained by the workflow engine, to sensor and telemetry data from the Fischertechnik components.
The following figure gives an overview of the main data sources and the kinds of data they hold:
5.2 The MySQL data model
The following diagram shows the concrete data model of the MySQL database, including the entities and attributes of the current implementation:
The database serves as the central store for:
- Product master data — bicycle models and their components,
- Customer data and delivery addresses,
- Stock levels for components and finished goods,
- Orders — customer orders, production orders, purchase orders — together with their status history,
- Vendor data.
5.3 Sample and master data
Documentation and sample data of the Fischertechnik factory are available in the Sample_data folder of the documentation repository. This data serves as:
- initial master data when setting up a new instance of the BPA Lab,
- test data for developing new components,
- a demonstration baseline for teaching sessions and presentations.
5.4 Process data and the foundation for process mining
In addition to persistent master data, the running processes generate substantial amounts of process data:
- Camunda 8 stores process instance data, variables, tokens, and lifecycle events.
- MQTT messages record the communication between job workers and controllers.
- Job worker logs capture technical operations and exceptions.
This event data is the foundation for process mining — a central teaching and demonstration component of the BPA Lab. Process mining allows students to:
- discover the actual flow of processes from event logs (and compare them to the modelled flow),
- check conformance between modelled and actual behaviour,
- analyse performance — durations, bottlenecks, rework loops.