The sales team of Nordlicht Reisen GmbH wants a report from bookings and web telemetry every morning, and on campaign days a dashboard that shows the load on the portal almost immediately. The booking database runs in Hamburg, the portal produces events, and hotels deliver files every day. This lesson chooses services: what brings the data together (integration) and where it is analyzed (analytics).
Prior knowledge. In module 4.13, K4 teaches the concepts: OLTP versus OLAP and the star schema (lesson 4.13.1), lakehouse and warehouse in Fabric including mirroring (4.13.2), ETL and ELT, watermarks and copy jobs (4.13.3), Spark (4.13.4) and the cost per office (4.13.5). None of that is repeated here. What is new is the choice between the services, the older ones included: when Azure Data Factory, when Fabric, when streaming services, when Databricks.
The requirement
Volumes and deadlines are assumptions of the task.
| No. | Requirement | Source | Measure (assumption) |
|---|---|---|---|
| D1 | Load bookings from the database in Hamburg every night without putting load on the booking system | Sales | Report ready by 06:00 |
| D2 | Take over the files of hotels and airlines every day | Partner interface | CSV in Blob Storage, a few GB |
| D3 | Analyze portal events almost immediately on campaign days | Marketing | Dashboard at most one minute behind the portal |
| D4 | Eight analysts work in Power BI and with SQL, nobody with Spark | Sales | 8 people |
| D5 | Data stays in the EU, access only for authorized people | Data protection | Base assumption |