You attended talk: An introduction to event sourcing (short)

Imagine for a moment you work for a large online retailer specialising in household goods. One morning, the head of marketing comes to you and says "I've had this great idea we're going to send discount vouchers to anyone who's changed their address in the past 3 months; people who've recently moved are more likely to be buying new furniture. Could you retrieve a list of all these customers?" You explain to him your systems only store a customers current address and doesn't record when it was last changed, a new feature is added to the backlog and the head of marketing leaves a little disappointed.

What if you could build a system which was able to answer this kind of question without knowing it up front? A possible solution is to use event sourcing, being able to go back to any previous state of your data is just one advantage of using event sourcing in your application. Event sourcing is a total paradigm shift from the more traditional model of storing an application's current state and can appear to be a very unnatural way to think about and build a system. In this talk I'm going to show you examples of why event sourcing can be a superior model and how to go about building an event sourced system in PHP.

Transcript

1. You Attended Talk: An Introduction To Event Sourcing Christopher Riley PHP Day, May 2016 1
2. Introduction
3. A familiar model 2
4. Deleting data LineItem table id order id product code quantity 1 1 PK00001 3 2 1 PK00002 1 3 2 PK00002 1 4 3 PK00007 4 5 4 PK00003 2 6 4 PK00005 6 3
5. Deleting data LineItem table id order id product code quantity 1 1 PK00001 3 3 2 PK00002 1 4 3 PK00007 4 5 4 PK00003 2 6 4 PK00005 6 4
6. Deleting data LineItem table id order id product code quantity deleted 1 1 PK00001 3 0 2 1 PK00002 1 0 3 2 PK00002 1 0 4 3 PK00007 4 0 5 4 PK00003 2 0 6 4 PK00005 6 0 5
7. Deleting data LineItem table id order id product code quantity deleted 1 1 PK00001 3 0 2 1 PK00002 1 1 3 2 PK00002 1 0 4 3 PK00007 4 0 5 4 PK00003 2 0 6 4 PK00005 6 0 6
8. Editing data LineItem table id order id product code quantity deleted 1 1 PK00001 3 0 2 1 PK00002 1 1 3 2 PK00002 1 0 4 3 PK00007 4 0 5 4 PK00003 2 0 6 4 PK00005 6 0 7
9. Editing data LineItem table id order id product code quantity deleted 1 1 PK00001 2 0 2 1 PK00002 1 1 3 2 PK00002 1 0 4 3 PK00007 4 0 5 4 PK00003 2 0 6 4 PK00005 6 0 8
10. Editing data LineItem table id order id product code quantity deleted version 1 1 PK00001 3 0 1 2 1 PK00002 1 1 1 3 2 PK00002 1 0 1 4 3 PK00007 4 0 1 5 4 PK00003 2 0 1 6 4 PK00005 6 0 1 9
11. Editing data LineItem table id order id product code quantity deleted version 1 1 PK00001 3 0 1 2 1 PK00002 1 1 1 3 2 PK00002 1 0 1 4 3 PK00007 4 0 1 5 4 PK00003 2 0 1 6 4 PK00005 6 0 1 1 1 PK00001 2 0 2 10
12. A familiar model (again) 11
13. Event sourcing
14. Event sourcing example 12
15. Advantages
16. Technical advantages • Debugging 13
17. Technical advantages • Debugging • Scalabilty 14
18. Technical advantages • Debugging • Scalabilty • Auditing 15
19. Technical advantages • Debugging • Scalabilty • Auditing • Testing 16
20. Example test 17
21. Summary
22. Retrieving data
23. Performance Problems?
24. CQRS
25. MVC Application 18
26. CQRS Application 19
27. Technical advantages • Scalabilty 20
28. Technical advantages • Scalabilty • Performance 21
29. Technical advantages • Scalabilty • Performance • Multiple read models 22
30. CQRS + Event sourcing Application 23
31. Should I use Event sourcing? • Complex to implement 24
32. Should I use Event sourcing? • Complex to implement • Development overhead 25
33. Should I use Event sourcing? • Complex to implement • Development overhead • Less understanding 26
34. Should I use Event sourcing? • Complex to implement • Development overhead • Less understanding • Hard to retofit 27
35. Conclusions
36. Further reading • Greg Young • https://github.com/qandidate-labs/broadway • https://github.com/carnage/broadway-bowling 28
37. Thanks • @giveupalready • https://github.com/carnage • http://joind.in/talk/1d28c 29