Asynchronous Web Server (AWS)

📋 Overview

The Asynchronous Web Server (AWS) is a high-performance Linux-based web server designed to serve files efficiently by leveraging advanced I/O paradigms. The primary goal of this project is to minimize CPU overhead and memory latency through asynchronous operations and zero-copy data transfer.

Key Technical Features:

• I/O Multiplexing: Utilizes the epoll API for high-scale, event-driven management of multiple simultaneous client connections.
• Zero-Copying: Implements sendfile for static content, bypassing user-space data buffering to accelerate throughput.
• Asynchronous File I/O: Employs Linux AIO (io_setup, io_submit) for dynamic file processing, ensuring the server remains responsive during heavy disk operations.
• Non-blocking Sockets: All network operations are performed on non-blocking sockets to prevent thread starvation and maximize scalability.
• Connection State Machine: Each client session is governed by a dedicated state machine to track the lifecycle of HTTP requests and responses.

🏗️ Server Architecture

The server categorizes content based on its location within the AWS_DOCUMENT_ROOT directory:

1. Static Files (/static/):

• Designed for assets that require no post-processing.
• Mechanism: Handled via sendfile (Zero-copy) for maximum efficiency.

2. Dynamic Files (/dynamic/):

• Designed for files that theoretically require server-side processing.
• Mechanism: Read from disk using the Asynchronous API (AIO) and pushed to clients via non-blocking sockets.

HTTP Implementation:

• Implements a functional subset of the HTTP/1.1 protocol.
• Response Codes: 200 OK for successful retrievals and 404 Not Found for invalid paths.
• Parsing: Utilizes a callback-based http-parser to extract resource paths and headers efficiently.

📂 Project Structure

.
├── aws.c               # Core server implementation (epoll loop, connection logic)
├── aws.h               # Macros, data structures, and configuration (port, root dir)
├── http-parser/        # External HTTP parsing library
├── tests/              # Automated testing suite
└── Makefile            # Build instructions

🛠️ Installation & Testing

Prerequisites

• A Linux-based environment (Kernel support for epoll and eventfd is required).
• gcc compiler and make utility.

Compilation

Build the executable by running the following command in the root directory:

make

Running the Automated Suite

The testing suite validates server functionality, API usage (sendfile, epoll, io_submit), and monitors for memory leaks.

cd tests/
make check

To execute a specific test case (e.g., Test 31):

./_test/run_test.sh 31

⚙️ Technical Deep-Dive

Connection State Machine

To manage asynchronous events, each connection structure maintains a state:

• STATE_RECEIVING: Reading and parsing the incoming HTTP request.
• STATE_SENDING_HEADER: Constructing and transmitting the HTTP response header.
• STATE_SENDING_DATA: Streaming the file content (via sendfile or AIO).
• STATE_CLOSING: Releasing resources and terminating the socket connection.

Advanced Linux APIs Utilized:

• Multiplexing: epoll_create, epoll_ctl, epoll_wait.
• Zero-Copy: sendfile.
• Async I/O: io_setup, io_submit, io_getevents, eventfd.

---

📝 Performance & Scalability

By combining event-driven multiplexing with asynchronous disk access, this server mitigates the "C10k problem." It minimizes context switching and memory copies, making it significantly more efficient than traditional thread-per-connection models.

Developed by: Daria-Ioana Drăghici]

Project: Operating Systems - Advanced Asynchronous Web Server Implementation