Валентин Кононов «CQRS и Event Sourcing для Backend-систем»

CQRS является достаточно известным подходом, но из-за некоторой сложности, на первый взгляд, мало кто его применял в реальных проектах. Мы поговорим о теоретической составляющей данного подхода и примерах из жизни, где его применение оправдано. Зачастую данных подход хорошо сочетается с другим принципом — Event Sourcing, который мы тоже затронем.

Доклад будет посвящен описанию подходов и сфере применения. Специфических знаний конкретного языка программирования или фреймворков для посещения данного доклада не требуется. Знания применимы независимо от платформы. Выступление и презентация вдохновлены известным .NET тренером и автором многих книг — Дино Эспозито, воркшоп которого Валентин посетил в Минске.

Transcript

1. CQRS and Event Sourcing in backend systems Valentin Kononov
2. CQRS
3. CQRS - what does it mean? Seek (if) You Are Ass
4. CQRS Meaning • Command • Query • Responsibility • Segregation Presentation Layer (Views, Inputs, Controllers) Infrastructure Layer (Repository, ORM, DataBase) Application Layer (App Service) Domain Layer (Domain Service, Event Bus, Message Bus, ...) Query Handler Data Access DTO Command Query
5. CQRS Concept Command = workflow, that alters the current state of the system Query = workflow, that reports a view of the current state of the system
6. Money Transaction Sample • User performs money transfer • Most probably he has money • We can perform transaction and report success • In case of error - rollback later
7. CQRS Diagram Canonical Layered Architecture Presentation Layer Application Layer Domain Layer Infrastructure Layer CQRS Layered Architecture Presentation Layer Infrastructure Layer Application Layer Domain Layer Data Access + DTO
8. CQRS Benefits - Divide et Impera • Distinct Optimization • Scalability Potential - both technical and Teams • Simplified Design and Less complexity of changes
9. CQRS - In simple code sample private int x; public int increment_and_return_x() { lock x; x = x + 1; int x_copy = x; unlock x; return x_copy; } private int x; public int value() { return x; } void increment_x() { x = x + 1; } Regular Code CQRS Code
10. Sample Solution Schema - In Visual Studio you can setup projects like this App.CommandStack App.Infrastructure App.QueryStack APP.CQRS
11. Event Sourcing
12. Event Sourcing • In the real world we observe events • In software, we tend to write models
13. Event Sourcing - All Changes to the Application are saved as a sequence of events - You don’t miss a thing. For the entire lifetime of the system
14. In some cases events are real data App State Event Storage Data
15. Event Sourcing Schema Application Write Interface Event Handler Event Store Application State Event Queue Read Service Application Read Interface
16. Read Model Artificially created data model that offers a business-specific view of the current state of the system. Event 1 Event 2 Event ... Event N Snapshot Event N+1 Event N+2 Event N+3
17. Lower abstraction layer of data • Events are immutable. You don’t miss a thing • Events can be replicated • Replaying the event doesn’t require to repeat the behavior • Track everything that happened
18. About Scalability • Event storage allows to implement any amount of new features and services based on the same data • Consistency - minimal set of rules allows to keep your data consistent • Can serve as environment for other platforms • Works well with microservices architecture
19. Technically - how to save events NOSQL Data Storage Special Event Store JSON Friendly Relational DB Raven DB Mongo DB Cosmos DB ... EventStore NEventStore SQL Server 2016 PostgreSQL ... Backendless Stores FireBase Parse ...
20. CQRS + Event Sourcing CQRS Event Based Applications Common Applications
21. Event Sourcing together with CQRS Command Pushed Workflow Execution Event Store Application State App State Update Event Saved New Event Triggered API Call Message Generated Notification Sent ...
22. Real Life samples of CQRS applicability And event sourcing as well
23. Projects based on these approaches 1. Sample of money transfer for CQRS 2. Sample of Social Networks 3. Sample of Trucks Telematic project with event sourcing 4. Sample of Reporting for high load once a year for CQRS 5. Sample of Electric Station for event sourcing
24. Social Network Sample • When user posts something - app produces number of work items ◦ Add data to event storage ◦ Add view storage for user ◦ Notify users (in browsers, push notifications….) ◦ Update search indexes... Web Application Application Service Search DataBase ... Cache Jobs Message Queue ... ...
25. Trucks Telematics • Eco System based on real time data from trucks ◦ Speed, location, temperature, … ◦ Engine turned on / off, … ◦ Diagnostic messages • Dashboard with average statistics and ability to change level of details • Reporting Web Dashboard Read Service Data Aggregation Event Store Truck Sensor Event Write Service
26. • System to enter finance data and see reports for government • Data is entered all the time, but at the end of finance year application load increases 10 times • Reporting should be available immediately Finance Year Reporting Reporting Reporting Service SQL Server Application Service Web Application MongoDB
27. • Real app data - events from turbines, solar panels. Each impulse of data - event. • No user interface to enter data • No presentation layer to observe data • Workers need to see some data model Electric Station
28. Q&A Speaker: Valentin Kononov Email: valentin.kononov@gmail.com Skype: valentin_kononov Thanks! Let’s discuss questions
29. Links • https://dev.by/lenta/oxagile/intervyu-s-avtorom-programming-microsoft-asp-net-mvc-dino-esposito • http://neventstore.org/ • https://habr.com/post/146429/ • https://habr.com/post/149464/