Skip to content

adr-03-backend-development-framework-selection.md

Latest commit

 

History

History
55 lines (38 loc) · 3.07 KB

adr-03-backend-development-framework-selection.md

File metadata and controls

55 lines (38 loc) · 3.07 KB

ADR-03: Backend Development Framework Selection

Status

Adopted

Context

Our application requires a robust backend framework that can handle international traffic, provide quick response times, and facilitate rapid development and deployment. We considered three popular frameworks across different programming languages: Node.js (Express), Java (Spring Boot), and Python (Django/Flask).

Node.js (Express):

  • Pros:
    • Performance: Event-driven, non-blocking I/O makes it suitable for I/O bound tasks and can handle a large number of simultaneous connections with low overhead.
    • Unified Stack: Uses JavaScript, allowing for a consistent language between the frontend and backend.
    • Ecosystem: Extensive package library (npm) and a vibrant community that contributes to a rich set of middleware and tools.
    • Development Speed: Facilitates rapid development due to its minimalist and flexible structure.
  • Cons:
    • CPU-Intensive Tasks: Not ideal for heavy computation tasks, though this isn’t our primary use-case.

Java (Spring Boot):

  • Pros:
    • Maturity: Battle-tested in numerous production environments; trusted for enterprise-level applications.
    • Performance: Robust performance, especially for compute-intensive tasks.
    • Ecosystem: Mature ecosystem with a wide array of libraries and tools.
  • Cons:
    • Development Speed: Typically requires more boilerplate and configuration than some newer frameworks.
    • Resource Overhead: Can be more resource-intensive compared to lightweight frameworks.

Python (Django/Flask):

  • Pros:
    • Rapid Development: Especially with Django, which follows the "batteries-included" philosophy, offering a comprehensive set of features out of the box.
    • Versatility: Suitable for both small applications (using Flask) and large applications (using Django).
    • Readability: Python’s syntax is clear and concise, leading to easily maintainable code.
  • Cons:
    • Performance: Generally slower than Node.js and Java, especially for I/O bound tasks.
    • Concurrency: Python's Global Interpreter Lock (GIL) can be a limitation for CPU-bound tasks, although there are workarounds.

Decision

Given our priority to maintain a streamlined tech stack and to achieve synergy between the backend and frontend, we have chosen Node.js with Express. This decision promotes the use of a single language (JavaScript) across our stack. Node.js with Express offers both performance suitable for our needs and rapid development capabilities.

Consequences

Pros:

  • Unified Development: Allows developers to leverage their JavaScript expertise across the entire application.
  • Fast Iteration: Rapid development and deployment cycles can be expected.
  • Scalability: Built to handle concurrent requests efficiently, making it suitable for our anticipated user growth.

Cons:

  • Complex Computation: For future needs, if heavy computation becomes a requirement, we might need to offload such tasks to dedicated services or consider integrating other languages/frameworks.