A TOY HTTP SERVER

This is my toy implementation of an extremely basic http server. Any and all suggestions are welcome.

High Level Design Targets

1. Core Server Layer

   • Main Listener: Accepts TCP connections.
   • Concurrency: Start with a simple thread-per-connection or thread pool; upgrade to async/event-driven I/O (later).
   • SSL/TLS: Add support for HTTPS (later).

2. Request Processing Pipeline

   • Parser: Parse incoming HTTP requests into struct, upgrade to a hash map. (later)
   • Router: Simple URL mapping to resolve static file paths.
   • File I/O: Read files from disk with MIME type detection.
   • Encoder: Build and encode HTTP responses.

3. Cross-Cutting Concerns (Low-Hanging Fruit)

   • Logging & Monitoring: Log requests, errors, and performance metrics.
   • Error Handling: Return proper 404/500 responses.
   • Security: Prevent path traversal and implement basic access control.
   • Compression: Support gzip/Brotli to reduce payload sizes.
   • Caching: In-memory LRU cache for frequently accessed files.
   • Directory Indexing: Support for index files or autoindex as needed.

4. Future Enhancements

   • Rate Limiting: Mitigate abuse and DoS attacks.
   • CDN Integration: Offload and globally cache static assets.

References

• Man Pages
• Beej's Guide
• The Makefile Tutorial
• The Linux Programming Interface