CQRS & Event Sourcing

Command-Query Responsibility Segregation by Stefan van Raaphorst

Transcript

1. d i g i t a l c r a f t s m a n s h i p CQRS
   and
   Event sourcing ! ! stefan van Raaphorst
2. Disclaimer - the experts • Greg Young - http://goodenoughsoftware.net/ • Udi Dahan - http://www.udidahan.com/ (The Software Simplist)
3. Disclaimer by the experts • NO it's not a silver bullet nor a Golden Hammer • Yes a lot of best practices and well known patterns are involved
4. CQS - Command-query separation Definition Separate command methods that change state from query methods that read state ! ! A principle Source: http://en.wikipedia.org/wiki/Command-query_separation [Bertrand Meyer]
5. Command-Query Responsibility Segregation Defined Definition Separate command messages that change state from query messages that read state. ! As Design - Architecture - Pattern from Domain Driven Design Source: http://codebetter.com/gregyoung/2009/08/13/command-query-separation/
6. Single model ! ! Source: http://martinfowler.com/bliki/CQRS.html
7. CQRS model Mindset: Eventual Consistency Data is out-of-date/stale the moment a user sees it on the screen in a collaborative environment.
8. Write and Read model Mindset This brings the notion of write models are different to read models and querying "search engine style" is a whole separate thing.
9. Commands and events Mindset: Intent in action Are you correcting an address or do you move someone to a different address
10. Event sourcing State History
11. Event sourcing Capture meta-data Intent - "Update" address was this Correcting a address or did this person move (so we can do some marketing) ! History/behavior - Shopping cart with two items or added 3 items and removed one
12. Event sourcing Mindset: Creating new reports and views for historical data
13. So what is the use? • Capture business changes in a lossless manner • Audit trail • It makes it easy to test, replay (business logic) failures in from the field
14. What are the benefits? • The ability to partition on a domain level and keep the model simple • Optimize for a single model (read/write) • To have choice in technology used (e.g. not one DB to tackle all issues)
15. d i g i t a l c r a f t s m a n s h i p ! ! stefan van Raaphorst http://www.thedeveloperday.com/tag/cqrs/