Fighting Bottlenecks with CQRS

More at http://verraes.net/ or http://twitter.com/mathiasverraes

How can you sell a gazillion tickets in an hour? This talk will explain how to use CQRS, Event Sourcing, and a strong focus on the problem domain to build a system that deals with peaks, reporting, and occasionally connected devices.

ResearchGate DevDay
Berlin, November 30, 2013

Transcript

1. Fighting
   Bottlenecks
   
   with CQRS/ES
   ResearchGate
   
   Berlin
   
   Nov 30, 2013
   Mathias Verraes
   
   @mathiasverraes
   
   verraes.net
   

   View Slide

2. Student
   
   of
   
   Systems
   Meddler
   
   of
   
   Models
   Labourer
   
   of
   
   Legacy
   

   View Slide

3. Blog
   
   verraes.net
   DDDBE
   
   domaindriven.be
   Podcast
   
   elephantintheroom.io
   

   View Slide

4. Business Problem
   
   A Simplified Model
   
   Command/Query Responsibility Segregation
   
   Event Sourcing
   
   Read Model Projections
   
   Eventual Consistency
   

   View Slide

5. Business Problem
   

   View Slide

6. Sell a gazillion concert tickets
   
   to desperate fans
   … in a hurry!
   

   View Slide

7. Don’t crash
   
   during peaks
   Don’t annoy
   
   our customers
   Don’t sell more
   
   than we have
   

   View Slide

8. A Simplified Model
   

   View Slide

9. Model invariants as a
   
   Presale Aggregate?
   

   View Slide

10. #reservedTickets #soldTickets
    #availableTickets
    Presale # Tickets
    

    View Slide

11. Model invariants as
    
    individual Ticket Aggregates
    

    View Slide

12. reserved sold
    available
    Individual Ticket status
    

    View Slide

13. Command/Query
    
    Responsibility Segregation
    

    View Slide

14. Client
    Remote Façade
    Application Services
    Domain Model
    ORM
    Database
    

    View Slide

15. Highly cohesive
    
    Read + Write
    Model?
    

    View Slide

16. Bottleneck
    

    View Slide

17. Client
    Database Database
    Remote Façade
    Application Services
    Domain Model
    ORM
    Eventually Consistent
    

    View Slide

18. CQRS challenges the assumption
    
    that reading and writing are
    
    supposed to share the same
    abstractions.
    

    View Slide

19. CQRS challenges the assumption
    
    that reading and writing are
    
    supposed to share the same
    models.
    

    View Slide

20. CQRS challenges the assumption
    
    that reading and writing are
    
    supposed to share the same
    databases.
    

    View Slide

21. CQRS challenges the assumption
    
    that reading and writing are
    
    supposed to share the same
    applications.
    

    View Slide

22. Segregation of read and write
    
    is a radical form of decoupling.
    

    View Slide

23. Client
    Domain Model
    Remote Façade Remote Façade
    Application Svcs Application Svcs
    Domain Model
    ORM ORM
    Database Database
    

    View Slide

24. Radical!
    

    View Slide

25. Client
    Write
    Model
    Read
    Model
    

    View Slide

26. Client
    Write
    Model
    Read
    Model
    DTO
    Commands
    Events
    

    View Slide

27. Interact
    
    with State
    Guard
    
    Invariants
    Project
    State
    

    View Slide

28. Task Based
    UI
    Event
    Store
    Polyglot
    
    Persistence
    

    View Slide

29. View Slide

30. Event Sourcing
    

    View Slide

31. TicketWasReserved [ ticketId, forUser, onDate ]
    

    View Slide

32. Ticket state = Ticket history
    
    !
    TicketWasSold [ toUser ]
    
    TicketWasReserved [ forUser, onDate ]
    
    ReservationWasReleased [ ]
    
    TicketWasReserved [ forUser, onDate ]
    
    TicketWasPrinted [ inPresale ]
    

    View Slide

33. Mature domains already work this way
    
    !
    AccountWasDebited [ 20 ]
    
    AccountWasDebited [ 10 ]
    
    AccountWasCredited [ 100 ]
    
    AccountWasOpened
    

    View Slide

34. BankAccount
    
    {
    
    debit(amount) {
    
    guardThatAccountIsNotOverdraft(amount)
    
    recordThat(
    
    new AccountWasDebited(amount)
    
    )
    
    }
    
    }
    

    View Slide

35. BankAccount
    
    {
    
    apply(AccountWasDebited event) {
    
    amount = amount - event.amount
    
    }
    
    }
    

    View Slide

36. BankAccount
    
    {
    
    getRecordedEvents() {}
    
    reconstituteFromHistory(eventHistory) {}
    
    }
    

    View Slide

37. Read Model Projections
    

    View Slide

38. TicketWasSold
    
    TicketWasReserved
    
    ReservationWasReleased
    
    TicketWasReserved
    
    TicketWasPrinted
    Ticket
    
    !
    !
    

    View Slide

39. TicketWasSold
    
    TicketWasReserved
    
    ReservationWasReleased
    
    TicketWasReserved
    
    TicketWasPrinted
    AvailableTicket
    
    TotalNumberOfAvailableTickets
    
    AverageNumberOfTicketsPerBuyer
    
    AverageNumberOfReleasesPerTicket
    

    View Slide

40. Eventual Consistency
    

    View Slide

41. Writes are immediately consistent
    
    Reads are eventually consistent
    

    View Slide

42. Many issues stemming from eventual consistency
    
    can be solved in the user interface
    

    View Slide

43. A Ticket can only be sold once.
    
    How do we know it’s available?
    

    View Slide

44. Get a batch of x*2 PossiblyAvailableTickets
    
    !
    ReserveTicket
    
    ack/nack
    
    loop while acks < 3
    To buy x tickets:
    

    View Slide

45. More?
    
    !
    Testing
    
    Context Mapping
    
    Event Storming
    
    Process Managers
    
    Occasionally Connected Systems
    
    Reporting
    
    Legacy Integrations
    
    …
    

    View Slide

46. Questions?
    verraes.net
    @mathiasverraes
    

    View Slide