Dan Nastasa

One of the most common approaches when designing loosely-coupled (micro)service architectures is using a database per service. But given the fact that these services are communicating asynchronously, how can we make sure that the data remains consistent between all our systems?

Change Data Capture gives us a solution to this challenge: it allows us to react to database changes (insert/update/delete operations) of interest and propagate specific events to our event bus.

During this masterclass, we will be addressing the following topics:

●     how to implement this pattern using open-source tools;

●     advantages and disadvantages of such an approach;

●     different ways to further improve our solution;

●     other areas where CDC might prove useful.

While this session will have its examples and coding done in Java, the topics and challenges presented here are generally applicable to distributed systems regardless of the language of implementation. Minimal knowledge of microservice architectural patterns and synchronous / asynchronous communication would ensure your best experience.  

Many times we find ourselves in the need of reacting to database changes and taking action depending on each case. While the use of triggers might be good enough initially, they tend to slow down the entire system when their number increases. But what about the binary log, a feature which sometimes people tend to leave behind?