Software Architecture Design

Proposed patterns

• Composite root - wires up the dependency injection container at startup
• Transaction Script - Organizes business logic by procedures where each procedure handles a single request from the presentation layer
• Domain Model - classes and structs interconnected to model the business domain and incorporates both behavior and data.
• Table Data Gateway - An object that acts as a gateway to a database table. One instance handles all the rows in the table.
• Row Data Gateway - An object that acts as a gateway to a single record in a data source. There is one instance per row.
• Repository Pattern - creates an abstraction layer between an application's business logic and its data storage. It provides a standardized, collection-like interface to query, add, and remove data, hiding the low-level database implementation details from the rest of the application.

Which architecture should I pick?

Summary

Scenario	Architecture
Small/simple app	Layered
Enterprise system with complex logic	Onion
Fast MVP/prototype	Layered
Long-term scalable platform	Onion
Mostly CRUD operations	Layered

Layered Architecture for Small-Medium Projects

Characteristics

• simple business logic that does not grow complex.
• You want fast development
• simple app, API, POC or MVP
• You are building:

• CRUD apps
• Internal enterprise systems
• Admin portals

Layered Architecture

Frontend

Presentation Layer - WebUI, Desktop App, Mobile App, CLI, Foundry Agent Service, more...

Backend

• Application (API) Layer
• Business Logic (Domain) Layer
  • Transaction Script pattern
• Data Access Layer
  • Table Data Gateway
  • Row Data Gateway
  • Repository pattern

Onion Architecture for Medium-Large Projects

Characteristics

• Complex business logic
• Need for long-term maintainability
• Multiple external dependencies (DB, APIs, queues)
• Requirement for easy testing
• Need to keep business rules independent from frameworks/technology
• You are building:

• Enterprise systems
• Banking/payment systems
• Large AI or logistics platforms
• Systems expected to evolve for years

Logging

API layer logs:

request received correlation ID / trace ID authenticated user / tenant response status unhandled exception

Application layer logs:

use case started important decisions or branches use case completed business failure that matters operationally

Domain layer logs:

generally no logging maybe only in rare cases if your domain services are acting like transaction scripts and you really need observability even then, keep it minimal

Infrastructure layer logs:

external calls failed DB retries / timeouts queue publish failures HTTP client failures adapter-specific technical details