Files
actions/README.md
Michał Michaluk 8709a957cf draft of readme
2017-12-21 17:56:15 +01:00

4.3 KiB

Missing complete example of Domain-Driven Design enterprise application

Command Query CRUD Responsibility Segregation

Not every piece of software is equally important... Not every piece will decide about company / product success or can cause not reversible negative business consequences like materialise brand risk or money loses. On the other hand scalability or non functional requirements are different for different activities in software. To accommodate to those differences, separate architectural patterns are applied:

Visualisation of Command Query CRUD Responsibility Segregation

Simple Create Read Update Delete functionality exposed with leverage of CRUD framework.

Goals of that approach:

  • fast initial development,
  • fast respond to typical changes (ex. „please add another 2 fields on UI”),
  • exposure of high quality API.

Complex Commands (business processing) going through Domain Model embedded in hexagonal architecture

Goals of that approach:

  • enable Implementing Domain Model first approach, by adding infrastructure adapters later,
  • keeping Domain Model as simple as possible by protecting it from accidental complexity caused by technological choices or transport models from external services,
  • make core business code technology agnostic, enabling continues technology migration and keeping long living projects up to date with fast evolving frameworks and libraries.

Complex Query made access from consumer facade to persistent data as direct and simple as possible:

  • query persisted projection of domain event in form of read model expected by consumer,
  • read model composed at query execution time build directly from persistent form of Domain Model,
  • mix of above: read model composed at query execution time build from pre-calculated persisted projection of domain event.

Additional complex calculation or projections can be partially delegated to Domain Model if desired.

Goals of that approach:

  • encapsulation of Domain Model complexity by providing consumer driven or published language API,
  • freeing Domain Model from exposing data for reads making Domain Model simpler,
  • improves reads performance and enable horizontal scalability.

Hexagonal Architecture

Only the most valuable part of that enterprise software is embedded in hexagonal architecture - complex business processing modeled in form of Domain Model.

Application Services - providing entry point to Domain Model functionality, Application Services are ports for Primary / Driving Adapters.

Domain Model - Object Oriented (in that case) piece of software modeling business rules, invariants, calculations and processing variants. Thanks to hexagon can be as clean and simple as possible - separating essential complexity of pure business from accidental complexity of technical choices, free of technical and convention constraints.

Ports - contract defined by Domain Model expressing expectations from external resources (services, database or other models). Declared interfaces alongside with IN-OUT parameters are Ports for Secondary / Driven Adapters.

Adapters - integration of technology (REST, database, external services, etc.) with Domain Model. Making useful application from Domain Model and technology.

Implementing Domain Model as first

In most projects the biggest risk is lack of domain knowledge among developers. We all known Java, databases and bunch of handy frameworks, but what about: Investment Banking, Automotive Manufacturing or even e-Commerce.

Lets face those risk at first, maintain and explore domain knowledge with Model Exploration Whirlpool and build Ubiquitous Language with your executable Domain Model from day one. Adding infrastructure and technology later is easy thanks to Hexagonal Architecture.

Starting from ZERO business knowledge through initial domain and opportunity exploration with Big Picture Event Storming:

Looking for system boundaries, impacted and required actors and there interactions with system under design:

Estimating depth of domain model and Command Query CRUD segregation:

Design level Event Storming with Domain Stories and Specification by Examples: <adjust-demand.feature>