Backend Benchmark

A fully automated benchmarking suite comparing popular backend frameworks (Python, Go, Node, Deno, Bun). It measures performance on a consistent JSON-processing route and outputs a standalone HTML dashboard with visual results.

⚡ What It Does

• Benchmarks dozens of frameworks across languages
• Measures Requests/sec, Latency, Transfer/sec
• Generates a zero-dependency HTML dashboard with charts + system info (results_dashboard.html)

📦 Frameworks Covered

• Python: Flask, FastAPI
• JavaScript Runtimes:
  • Node: Node (native), Express, Fastify, NestJS (Express), NestJS (Fastify)
  • Bun: Bun (native), Express, Hono, Elysia
  • Deno: Deno (native), Express, Hono
• Go: Gin, Echo, Fiber, Native net/http

🔬 Benchmark Scenario

Route: POST /process

• Input JSON: { "numbers": [1, 2, 3, 4, 5] }
• Task: Compute the sum of squares of the numbers
• Output JSON: { "result": 55 }

Each framework is tested using:

wrk -t"$CORES" -c1000 -d60s -s post.lua http://localhost:3000/process

📁 Project Structure

backend-benchmark/
├── <framework>/         # One folder per framework
│   └── install.sh       # Install dependencies of the framework (optional)
│   └── start.sh         # Starts that framework's server
├── results/             # Results ouput folder
│   └── raw/             # Raw wrk output per framework
│   └── results_summary.csv  # Parsed performance data
│   └── results_dashboard.html  # Interactive HTML report
├── post.lua             # wrk load script
├── benchmark_presetup.sh  # Install all dependencies
├── benchmark_runner.sh  # Run benchmarks
└── parse_html.py        # Parses wrk results into charts

🔧 Prerequisites

Install the following tools:

• Python 3.11+
• Node.js (LTS)
• Go
• Deno
• Bun
• wrk

Ensure each tool is available in your PATH.

🚀 Usage

1. Clone & enter project

git clone https://github.com/Drarox/Backend-Benchmark.git
cd backend-benchmark

2. Run presetup (install dependencies per framework)

./benchmark_presetup.sh

3. Run benchmarks (fully automated)

./benchmark_runner.sh

This will:

• Sequentially start each framework server
• Run a high-load test using wrk
• Kill the server
• Save results in results/
• Generate results_dashboard.html

🐳 Run with Docker (No Host Dependencies)

Don't want to install Python, Node, Go, Deno, Bun, or wrk on your machine? No problem: everything runs cleanly inside a container. The Docker image installs all the prerequisites !

✅ One-Time: Build the Docker image

docker build -t backend-benchmark .

🚀 Run the full benchmark suite

docker run --rm -v "$PWD/results:/app/results" backend-benchmark

• Benchmarks all frameworks
• Generates the HTML dashboard
• Mounts results to your host in the results/ folder

This lets you run the full suite with zero host setup and clean everything up with one docker rmi.

🧾 Benchmark Results

• macOS – M1 Pro (8-core), 16 GB RAM — View dashboard (May 28, 2025)

• macOS (Docker Desktop) – M1 Pro (8-core), 8 GB RAM — View dashboard (May 28, 2025)

• Ubuntu (Docker) – Xeon D-1531 (6-core), 32 GB RAM — View dashboard (May 28, 2025)

• Ubuntu – Xeon D-1531 (6-core), 32 GB RAM — View dashboard (May 30, 2025)

• Ubuntu (Docker) – AMD Ryzen 7 1700X (8-core), 24 GB RAM — View dashboard (May 30, 2025)

🤝 Contributing

Pull requests welcome! Add frameworks, improve charts, or enhance the automation. Feel free to fork this repository and submit a pull request with your changes.

📄 License

MIT — use freely, modify openly, benchmark responsibly.