WebSocket Comparison Project

This project aims to compare WebSocket implementations in two Python frameworks: Tornado and Websockets. The evaluation includes measuring latency, throughput, scalability, and collecting hardware details during data exchange.

Prerequisites

1. Clone the repository: git clone https://github.com/Rectiras/repo-websockets.git
2. Navigate to the project directory: cd repo-websockets
3. Ensure Python is installed on your system.
4. Install project dependencies: pip install -r requirements.txt

Tornado WebSocket Implementation

Files for Tornado

1. tornado_client.py: Tornado WebSocket client. Sends messages to the server and generates an Excel file named tornado.xlsx containing real-time data, including latency, throughput, and hardware details.

2. tornado_server.py: Tornado WebSocket server. Listens for incoming messages from the client, simulates real-time data, and calculates latency and throughput.

Running the Tornado WebSocket Project

1. Start the server: python tornado_server.py
2. Launch the client: python tornado_client.py

Using Tornado WebSockets

• The client can input messages and receive responses from the server with fixed JSON data (modifiable in tornado_server.py).
• To generate the Excel file with a summary, clients must input exit.

Websockets WebSocket Implementation

Files for Websockets

1. websockets_client.py: Websockets WebSocket client. Sends messages to the server and generates an Excel file named websockets.xlsx containing real-time data, including latency, throughput, and hardware details.

2. websockets_server.py: Websockets WebSocket server. Listens for incoming messages from the client, simulates real-time data, and calculates latency and throughput.

Running the Websockets WebSocket Project

1. Start the server: python websockets_server.py
2. Launch the client: python websockets_client.py

Using Websockets

• Similar to Tornado, the client can input messages and receive responses from the server with fixed JSON data (modifiable in websockets_server.py).
• To generate the Excel file with a summary, clients must input exit.

Results

1. Latency Comparison

   • Tornado Average Latency: 0.000965

   • Websockets Average Latency: 0.000876

   • Websockets Performed 10% better than Tornado

2. Throughput Comparison

   • Tornado Throughput: 30,2804844366243

   • Websockets Throughput: 30,3015221337059

   • Almost Identical Throughput with Websockets performing slighly better than Tornado.

3. Scalability

   • Tornado:

   • • Well-suited for handling concurrent connections
   • • May require fine-tuning for specific scalability requirements

   • Websockets:

   • • High concurrency support
   • • Asynchronous architecture for efficiency

Conclusion

Based on the comparison, Websockets demonstrated approximately 10% better latency and nearly identical throughput compared to Tornado. While both libraries have their strengths, Websockets is recommended for this scenario due to superior latency performance.