🌐 Multi-threaded Web Server with Synchronization and Scheduling

📝 Overview

This project is an implementation of a multi-threaded web server in C, completed for the Technion's "Web Servers and Synchronization" course (HW3). The server handles concurrent HTTP GET requests and serves both static 🗂️ and dynamic ⚙️ content. Starting from a basic single-threaded web server, the project adds advanced features like thread pooling, request queuing, scheduling algorithms, and usage statistics 📊.

---

🚀 Features Implemented

🔁 1. Multi-threading

• Transformed the original server into a multi-threaded version using POSIX threads 🧵.
• Created a thread pool that handles incoming HTTP requests concurrently.
• The main thread accepts connections and enqueues them 📬.
• Worker threads dequeue and process requests 🛠️.

🧩 2. Synchronization

• Used pthread_mutex for locking 🔒 and pthread_cond for signaling 🔔.
• Prevented race conditions and ensured thread-safe queue operations.
• Blocked producer/consumer threads properly without busy-waiting 🧘‍♂️.

⚖️ 3. Overload Scheduling Policies

Handled request overflow with different strategies:

• 🧱 block – Wait until space is available.
• ✂️ drop_tail (dt) – Drop the newest request.
• 🗑️ drop_head (dh) – Drop the oldest queued request.
• 🚫 block_flush (bf) – Wait until queue is empty, then drop.
• 🎲 drop_random (random) – (Bonus) Drop 50% of waiting requests randomly.

🧨 4. Special Suffix: .skip

• For requests ending in .skip, the server dequeues and defers handling the last request in the queue 🔄.
• Ensured correct queue accounting while applying the .skip logic.

📈 5. Usage Statistics

Embedded statistics in HTTP response headers:

🕒 Per Request:

• Stat-Req-Arrival – Time of arrival.
• Stat-Req-Dispatch – Wait time in queue.

🧵 Per Thread:

• Stat-Thread-Id – Thread index.
• Stat-Thread-Count – Total requests handled.
• Stat-Thread-Static – Static requests count.
• Stat-Thread-Dynamic – Dynamic requests count.

Headers are visible via custom clients or tools like curl 🛠️.

---

⚙️ How to Run

make
./server [port] [num_threads] [queue_size] [schedalg]