5+ years ago, Saxo introduced a new reference architecture mandating that all new service-to-service communication is implemented using a centralized Kafka data platform. Some of the objectives were
A highly available and scalable async / event driven interfaces rather than sync / batch connections. Centralized governance of data contracts, ownership, schema evolution and well documented contracts.
The investment now pays dividends in new ways – such as supporting AI with a robust, well documented, low-friction, real-time data foundation.
Key takeaways:
– Async event driven service-to-service communication can enable services, but also AI beyond classic batch data warehousing
– A robust governance model on the schemas used by the producers extends into analytical use cases and pays off for everyone
