What is Void ? - Click here to know
Building the User-Facing Side with Flask
Today, I’m starting on the user-facing side of the platform using Flask. Since this is still the MVP phase, I’m keeping things lightweight while focusing on getting the core features in place. Flask will handle user interactions, allowing them to submit their datasets and models. Once submitted, the datasets will be divided and distributed to the available nodes in the network.
To manage the job queues and distribute tasks to the nodes, I’ll be using Celery for asynchronous task handling, paired with Redis as the message broker and task queue backend. This setup will ensure efficient communication between the Flask web application and the nodes.
The nodes themselves won’t be running Flask; instead, they will be running Void Core, which I plan to build as a desktop app using Rust or C++ for high performance and efficient resource management. Each node will receive a portion of the dataset, process the model, and return the trained results back to the main platform, which will combine them to form a global model.
By keeping Flask isolated to the web interface and using Rust or C++ for the node-side application, I can ensure that the backend is optimized for performance while providing an easy-to-use interface for users. Flask, Celery, and Redis make it easier to manage job distribution, while Rust/C++ allows each node to train models efficiently in its own virtual environment.
This setup gives me the flexibility to expand the platform further while keeping things tidy and efficient. Let’s see how it unfolds.
Top comments (0)