Data Mesh promotes the idea of moving from centralised analytics that collects the necessary data from multiple upstream sources using complex data pipelines, to decentralised data products that are supported by domain-centric teams. At Event Store, we promote Event Sourcing as a way to build systems and persist state, where you don’t lose history. We found that domain events share both operational and analytical characteristics, and we’d like to share our ideas about building data products in event-sourced systems.
Topics we will cover:
- What’s Event Sourcing
- Message: domain events are persisted behaviour
- What’s Data Mesh
- What Data Mesh means in DDD terms. Analytics context being a conformist. Moving from conformist to customer-supplier
- Lambda architecture
- Query models in Event Sourcing, overview
- Query model as the batch layer source, isolating a dedicated query model to serve the data product
- Speed layer using domain events
- Kappa architecture
- Events - operational and analytical data at the same time, shared characteristics
- Can domain events be a data product on their own?
- What happens when the domain model changes and the events schema changes too?
- Optimise for your scale, you’re probably not Spotify or Uber
%20alt.svg){kind=small}