Distribute Your Microservices Data With Events, CQRS, and Event Sourcing

You split up your monolithic codebase into smaller artifacts and thought you were done! Now you’ve reached the hardest part: how do you split your data, and how do you keep your system working with it?

If this rings a bell, then you need to register for this session. We’ll explore how events with an event-driven architecture can help you succeed in this distributed data world. Learn concepts like CQRS, event sourcing, and how you can use them in a distributed architecture with REST, message brokers, and Apache Kafka.

Transcript

1. Distributed Your Microservices Data With Events, CQRS, and Event Sourcing Edson Yanaga @yanaga
2. @yanaga 2 @yanaga
3. @yanaga 3 bit.ly/mono2microdb
4. @yanaga 4 Code is easy, state is hard
5. @yanaga 5 HTTP + REST/gRPC/RMI/IIOP Synchronous Data Retrieval on a Remote Endpoint Client Canonical Source of Information
6. @yanaga 6 HTTP + REST/gRPC/RMI/IIOP Synchronous Data Retrieval on a Remote Endpoint Client Canonical Source of Information
7. @yanaga 7 Latency Availability Performance
8. @yanaga 8 Cache and/or Polling
9. @yanaga 9 Eventual Consistency
10. @yanaga 10 Eventual Consistency
11. @yanaga 11 First, a look back at the past…
12. @yanaga 12 How was data managed 10 years ago?
13. @yanaga 13 Terrified about Entity Beans
14. @yanaga 14 Hibernate to the rescue!
15. @yanaga 15 Replacing XML with @Annotations
16. @yanaga 16 POJOs as an (Anemic) Domain Model
17. @yanaga 17 Event Sourcing
18. @yanaga 18 ID CUSTOMER_ID BALANCE 1001 990 1000 1002 991 0 1003 991 -500 1004 992 300 Account
19. @yanaga 19 ID ACCOUNT_ID TIMESTAMP OP AMOUNT 1 1001 1234567890 C 1000 2 1002 1234567891 C 200 3 1001 1234567900 D 300 4 1001 1234567995 D 150 Transactions
20. @yanaga 20 Enables you to think in the Events that happened in the system
21. @yanaga 21 CQS (Command-Query Separation) “Asking a question should not change the answer” (Bertrand Meyer)
22. @yanaga 22 CQRS (Command Query Responsibility Segregation)
23. @yanaga 23 CQRS (Command Query Responsibility Segregation)
24. @yanaga 24 ID NAME PHONE ADDRESS BIRTH 1 Burr 222-222-2323 901 South St 12/12/1968 2 Edson 222-333-3434 112 North Dr 03/03/1978 3 John 111-456-4545 666 Iron St 06/06/1966 4 Doe 333-789-7890 777 Boeing Dr 07/07/1977 INSERT INTO CUSTOMER(ID,NAME,PHONE,ADDRESS,BIRTH); SELECT * FROM CUSTOMER;
25. @yanaga 25 SELECT ID, NAME, PHONE FROM CUSTOMER; SELECT ID, NAME, ADDRESS FROM CUSTOMER;
26. @yanaga 26 CQRS with separate data stores
27. @yanaga 27 SELECT ID, NAME, AGE, AVG_BILL FROM CUSTOMER_REPORT_VIEW; SELECT ID, PHONE, LAST_PAYMENT_AMOUNT FROM CUSTOMER_BILLING_VIEW;
28. @yanaga 28 CQRS & Event Sourcing
29. @yanaga 29 ID CUSTOMER_ID BALANCE 1001 990 1000 1002 991 0 1003 991 -500 1004 992 300 Account ID ACCOUNT_ID TIMESTAMP OP AMOUNT 1 1001 1234567890 C 1000 2 1002 1234567891 C 200 3 1001 1234567900 D 300 4 1001 1234567995 D 150 Transactions
30. @yanaga 30 ID CUSTOMER_ID BALANCE 1001 990 1000 1002 991 0 1003 991 -500 1004 992 300 Account ID ACCOUNT_ID TIMESTAMP OP AMOUNT 1 1001 1234567890 C 1000 2 1002 1234567891 C 200 3 1001 1234567900 D 300 4 1001 1234567995 D 150 Transactions READ MODEL WRITE MODEL
31. @yanaga 31 Why CQRS?
32. @yanaga 32 Performance
33. @yanaga 33 Now, Back to the Future!
34. @yanaga 34 Distribution Availability Integration Analytics
35. @yanaga 35 Canonical Source of Information (Write Data Store) µ
36. @yanaga 36 Canonical Source of Information (Write Data Store) Read Data Store Read Data Store µ µ µ
37. @yanaga 37 Canonical Source of Information (Write Data Store) Read Data Store Read Data Store µ µ µ Events Events
38. @yanaga 38 HTTP + REST/gRPC/RMI/IIOP Synchronous Event Retrieval Client Canonical Source of Information µ µ
39. @yanaga 39 HTTP + REST/gRPC/RMI/IIOP Synchronous Event Retrieval Client Canonical Source of Information µ µ
40. @yanaga 40 AMQP/Kafka Asynchronous Event Propagation Client Canonical Source of Information µ µ
41. @yanaga 41 Approaches
42. @yanaga 42 1st: You need change events (Event Sourcing)
43. @yanaga 43 Change your application OR Use Debezium
44. @yanaga 44 2nd: You need to distribute the events
45. @yanaga 45 Create a REST endpoint for your events using Quarkus Or Publish the events on ActiveMQ / Kafka
46. @yanaga 46 3rd: Consume the events and update your local CQRS data store
47. @yanaga 47 4th: Use the local data in your local CQRS data store
48. @yanaga 48 Thank you!