Clustered Event-Driven Services

Please log in to watch this conference skillscast.

In 2010 I came across a strange, new (to me!) architecture that the LMAX team used for their Foreign Exchange system. You might have heard about the Disruptor, it came out of this project.

The core of our system is a clustered service which uses the Raft consensus algorithm to reliably replicate state between the different nodes and hosts our application logic. We will take a quick look at Raft and then at the benefits of this design compared to more “mainstream” architectures. This architecture offers a clean separation of concerns between the infrastructure – which takes care of the concurrency, I/O and high availability aspects – and the application logic. The clean architecture is a great fit for domain-driven design.

If you fancy building fast, resilient services without a database you should come to this talk!

YOU MAY ALSO LIKE:

• Clustered Event-Driven Services (SkillsCast recorded in December 2017)
• The Science of Queues: Performance Monitoring for Themes Parks and Distributed Systems (SkillsCast recorded in December 2020)
• Solving Problems like a Game Designer (SkillsCast recorded in December 2020)