Luglio 4, 2024 Vincenzo Carletti

Lost password

This page provides detailed information about the dataset proposed in the paper”Enhancing IoT Network Security with Graph Neural Networks for Node Anomaly Detection“.
We extracted graphs from three IoT communication databases:

  1. IoT-23, for more information and original *.pcap file check the link
  2. IoTID20, for more information and original *.pcap file check the link
  3. IoTTraces, for more information and original *.pcap file check the link

If you want to extract the dataset yourself, please refer to the instructions in the linked repository

IoT-23

Starting from the *.pcap files, we extracted graphs for two representations: Traffic Dispersion Graph (TDG) and Extended-TDG (e-TDG).
Due to time and computational constraints, we were able to extract only a subset of the captures. A complete list of the extracted graphs is provided below.
Note: The results reported in the paper refer to the subset of extracted e-TDG graphs to ensure the results are comparable.

From the marked captures, we extracted the graphs considering different snapshot intervals. The statistics of the obtained dataset are provided below.

The dataset can be downloaded from this link.

IoTID-20

For the IoTID-20 database, we extracted graphs for all the *.pcap files, obtaining a dataset with the statistics reported below.

For each snapshot we show the total number of graphs (|G|) together with the average number of nodes (|V|}) and edges (|E|}) and the corresponding standard deviation.

The dataset can be downloaded from this link.

IoT Traces

For the IoT Traces database, we extracted graphs for all the *.pcap files, obtaining a dataset with the statistics reported below.

For each snapshot we show the total number of graphs (|G|) together with the average number of nodes (|V|}) and edges (|E|}) and the corresponding standard deviation.

The dataset can be downloaded from this link.