🎥 Dynamic Visualization of SST Simulator Traces

Simulating large-scale systems using tools like the Structural Simulation Toolkit (SST) generates a wealth of data about network behavior, communication patterns, and resource utilization. However, analyzing these traces can be challenging without intuitive, interactive visualizations. Leveraging the Network Traffic Visualizer [1], this project will create a dynamic visualization framework tailored to SST traces.

The goal of the project is to adapt and extend the Network Traffic Visualizer to visualize network traffic data generated by the SST simulator. The visualization will represent the network as a graph, where nodes correspond to routers or endpoints, and links dynamically change color and thickness over time to represent traffic intensity and bandwidth usage. This real-time visualization will help identify bottlenecks, load imbalances, and the impact of different routing algorithms or topologies.

The work will involve:

• Parsing and converting SST simulator trace outputs to a format compatible with the Network Traffic Visualizer tool.
• Integrating additional features into the visualizer, such as traffic filtering (e.g., by specific nodes or links) and support for large-scale networks.
• Benchmarking the tool’s scalability and performance when visualizing large SST simulations.
• Conducting case studies on SST-generated network topologies (e.g., Dragonfly, Fat-tree) to demonstrate the tool’s effectiveness.

The tool will serve as a critical resource for analyzing simulated networks, particularly for researchers designing and optimizing interconnects for high-performance computing.

Skills required:

• Knowledge of C/C++ and Python programming languages.

[1] Network Traffic Visualizer
[2] Structural Simulation Toolkit (SST)

Approximate composition: 15% State of the art analysis, 35% Tool Integration/Development, 50% Implementation/Experiments