Performance aspects of Axon-based CQRS/ES systems

Комментарий Программного комитета:

Рассказ о решениях проблем производительности CQRS/ES-систем от одного из авторов этих решений в популярном фреймворке — отличная возможность починить наконец свой production.

Axon Framework is a well-established open source framework for writing CQRS, event-sourcing (ES) and DDD-based applications in Java. It has been around for eight years and is used in business applications around the globe, with the financial services industry proving to be a sweet spot.

In October 2018 Axon 4 has been released. Starting from this release, Axon is a platform consisting of the open source Axon Framework and the open source Axon Server. The platform offers an unique, scalable way to build microservices. It allows developers to focus on functionality, with things like the correct routing of messages being handled by the platform.

In this session, we'll zoom into a specific aspect of Axon-based CQRS/ES systems: performance. While CQRS/ES is often touted as a great performance booster (through query simplification and eventual consistency), in reality, there are some challenges. Large event stores, or individual aggregates with large numbers of events, may show horrible performance. Also, replaying the event stream into a new read model sounds easy in principle, but becomes challenging when talking about billions of past events.

This session will begin with a fast-paced review of the core CQRS/ES principles and their implementation in Axon. After that, we'll focus on the various performance challenges and approaches to solving them.