Availability-Aware Routing

This repository contains implementations and evaluation scripts related to the following paper: Balázs Vass, Levente Birszki, Erika Bérczi-Kovács, Péter Babarczi, Péter Gyimesi, and János Tapolcai, "Availability-Aware Routing in the Presence of Geographically Correlated Failures," IEEE INFOCOM 2022 - IEEE Conference on Computer Communications, Tokyo, Japan, May 2026

---

Repository Structure

• /test_data
  Input graphs and SRLG (Shared Risk Link Group) data.

• /cpp_codes
  Black-box routing algorithms implemented in C++.

• /results
  Output files generated by running the experiments.

• /figures
  Plots and visualizations generated from the results.

• measures.ipynb
  Notebook used to run experiments and generate raw results.

• results.ipynb
  Notebook for analyzing results and creating figures.

• my_utils.py
  Helper and utility functions used across notebooks.

---

Dependencies

The following external Python packages are used for data processing, analysis, and visualization:

• networkx – graph representation and processing
• numpy – numerical computations
• pandas – result analysis and data handling
• plotly – plotting and visualization
• dash – interactive visualizations built on top of Plotly

---

Usage

Compile C++ Codes

The C++ implementations in /cpp_codes can be compiled using g++ as follows:

g++ dateline_capacity.cpp -o dateline_capacity -std=c++11
g++ dateline_prob.cpp -o dateline_prob -std=c++11

Run Experiments

After compiling the C++ binaries, all experiments can be re-run using measures.ipynb.

⚠️ Warning:
Running measures.ipynb will overwrite the contents of the /results directory.

Analyze Results

To analyze existing results and generate figures, run the appropriate cells in results.ipynb.