Data Driven: The Journey to Bring Real-time ERP Data to a Lakehouse


TRACKData Engineering and Streaming
SKILL LEVELIntermediate

For many years now, Michelin has been on the road to becoming more of a data-driven company every day. Last year, we gave a testimony on our wish to democratize data inside and outside of our group. To support this wish, the ERP of a company is one of the pillars that more than likely gathers the most interesting data to analyze when it comes to finance, purchasing, controlling, etc. That's why it is key to bring the ERP data on a datalake, embracing a Data Mesh architecture, to unleash the power of business users to perform analysis. But what if you broughtt the data in real-time? Have inventory transactions updated in less than a minute? This is a game changer. This testimony is the story behind the scenes when it comes to implementing a change data capture system on a 15 terabyte Oracle E-Business ERP with multiple instances, implementing tools such as Kafka, Qlik, Oracle Golden Gate, Databricks Delta Live Tables, Structured Streaming, Dremio and PowerBI.