dataset-graphscaling

Interactive version here

Structure

graph-generation:

• generated graphs
• divided by method:
  • Netgan
  • BiGG
  • GraphRNN
  • GSM
  • GraphGDP

graph-upscaling:

• up-scaled graphs
• divided by method and scale:
  • Kronecker:
    • scales 2-256
  • Gscaler:
    • scales 2-256
  • EvoGraph:
    • scales 2-256

Use the Data

All graphs are stored as a pickled list of networkx graphs.

You can load graphs using the following code:

import pickle
import networkx as nx

with open(filename, 'rb') as f:
    graphs = pickle.load(f)
    print(type(graphs[0]))

<class 'networkx.classes.graph.Graph'>

Cite

@inproceedings{Helm2310:Synthesizing,
  author = {Helm, Max and Carle, Georg},
  title = {{Synthesizing and Scaling WAN Topologies using Permutation-invariant Graph Generative Models}},
  booktitle = {19th International Conference on Network and Service Management (CNSM 2023)},
  address = {Niagara Falls, Canada},
  pages = 6,
  days = 30,
  month = oct,
  year = 2023,
  month_numeric = 10
}