This project implements a **Sockets** based distributed algorithm for communication between nodes. The code is designed to be containerised using Docker and executed locally. The algorithm will later be used as a backbone for distributed federated learning. This version only deals with message communication between neighbors: there is no training/learning. The messages will later represent the weight exchanges of local neural networks (NN) deployed on the nodes.
This project implements a **Sockets** based distributed algorithm for communication between nodes. The code is designed to be containerised using Docker and executed locally. The algorithm will later be used as a backbone for distributed federated learning.
This version only deals with message communication between neighbors: there is no training/learning. The messages will later represent the weight exchanges of local neural networks (NN) deployed on the nodes.
## Prerequisites
...
...
@@ -39,6 +41,10 @@ Where <num_nodes> is the number of nodes in your network.
Replace the varibale `image` (line 3) with your previouly built container image.
## Local execution
docker-compose up
## Troubleshooting
- Check the container logs
- Ensure that all environment variables are correctly set and that the neighbors' file is correctly configured.