Network Convergence with Gossip and Push-Sum Algorithms

Overview

This project explores the convergence times of Gossip and Push-Sum algorithms across different network topologies. We implemented failure models to simulate real-world conditions where nodes or connections might fail.

What is Working

• Gossip Algorithm: Successfully implemented and tested across Full, Line, 3D Grid, and Imperfect 3D Grid topologies.
• Push-Sum Algorithm: Successfully implemented and tested across Full, Line, 3D Grid, and Imperfect 3D Grid topologies.
• Graphical Analysis: Plots generated to visualize convergence times relative to network size.

Largest Network Handled

Full Topology:

• Gossip Algorithm: Managed up to 10000 nodes.
• Push-Sum Algorithm: Managed up to 10000 nodes.

Line Topology:

• Gossip Algorithm: Managed up to 10000 nodes.
• Push-Sum Algorithm: Managed up to 10000 nodes.

3D Grid Topology:

• Gossip Algorithm: Managed up to 10000 nodes.
• Push-Sum Algorithm: Managed up to 10000 nodes.

Imperfect 3D Grid Topology:

• Gossip Algorithm: Managed up to 10000 nodes.
• Push-Sum Algorithm: Managed up to 10000 nodes.

Graphical Results

These plots illustrate the convergence times for both algorithms across different network sizes and topologies, highlighting the scalability and performance differences.