Oliveyoung's CDC Transition for Real-Time Campaign Targeting
Key point
To solve the data latency problem of its existing ODI-based batch system, Oliveyoung built a hybrid CDC architecture leveraging OGG and Kafka.
Details
Oliveyoung's existing campaign system had been synchronizing data using a batch approach with ODI (Oracle Data Integrator). However, data latency ranging from 20 minutes to 1 hour created a limitation in that it could not immediately reflect customers' latest behavior.
This latency caused issues such as degrading customer experience by sending notifications to customers who had opted out of marketing, or placing excessive load on the source DB during large-scale data changes. To solve this, Oliveyoung decided to introduce a real-time data pipeline based on OGG (Oracle Golden Gate) and Kafka.
However, the OGG connector alone could not handle the existing complex data joins and transformation logic. Accordingly, Oliveyoung adopted the following hybrid CDC architecture.
- OGG Connector: Handles the publishing of large volumes of messages that require simple 1:1 mapping, ensuring stability.
- Campaign-Worker: Handles the processing and integration of data that requires joining multiple tables or complex business logic.
Through this structure, Oliveyoung improved data consistency and laid the foundation for real-time customer targeting.
This summary was generated automatically by AI. Check the original for the author's claims and context. Copyright belongs to the original author.
Our guide explains how the AI works. Report summary errors, attribution issues, or removal requests via Contact.