simple-http

Multi-threaded HTTP server built using only the Python standard library

Server instances compatible with reverse proxy

Features

• HTTP/1.1 subset
• Serves Static Files: Delivers HTML, CSS, JS, image, json, txt, and other data formats
• Support for IPv4 connections via TCP
• Handles multiple client requests simultaneously with multi-threading (thread-per-request model)
• HTTP Methods: Support for GET, POST (subset), OPTIONS, and HEAD methods
• HTTP Headers: Support for Content-Type, Content-Length, Connection, ETag, If-None-Match, Date, Last-Modified, Host, and User-Agent headers
• HTTP Responses: Support for 200 OK, 204 No Content, 304 Not Modified, 400 Bad Request, 403 Forbidden, 404 Not Found, 405 Method Not Allowed, 500 Internal Server Error, 501 Not Implemented, and 505 HTTP Version Not Supported messages
  • Correct error handling for path traversal vulnerabilities (403 Forbidden) and permissions for dynamic files (405 Method Not Allowed)
  • Correctly returns 304 Not Modified if strong ETag is provided and validated
• Persistent connections with keep-alive

Get Started

With Python 3.x

1. Clone this repository:

git clone https://github.com/davidmenggx/simple-http && cd simple-http

2. Run the server:

python3 server.py

Optional: Specify port number, keepalive time (seconds), and verbose
Defaults to port 8080, 5 seconds keep-alive, no logging

python3 server.py --port 5678 --keepalive 30 -v

Usage

Base directory is public/

Dynamic assets in api/ allow POST, GET, OPTIONS, HEAD
All other static assets allow GET, OPTIONS, HEAD

public/
├── api/
│   ├── comments.json
│   └── public_file.txt
├── data /
│   ├── users.json
│   └── videos.json
├── images /
│   └── img1.jpg
├── index.html
├── script.js
└── styles.css

Testing

Fetch index.html (using default port 8080):

curl http://localhost:8080/

Visit the front page: http://localhost:8080 (or the port you specified)


Make a post to api/public_file.txt:

curl -i -X POST -d "Hello from GitHub" http://localhost:8080/api/public_file.txt

Now read api/public_file.txt:

curl -i http://localhost:8080/api/public_file.txt

Performance

Using Locust*, one server instance was able to achieve up to 2500 RPS, 40 ms median latency, and 80 ms p99 latency with 100 concurrent connections using persistent connections**

*Running on a Dockerized environment on my laptop (8-core Intel Ultra 7 with 32 GB of RAM on Windows 11)

**For reference, without persistent connections I am only able to achieve 1000 RPS