14 categories of encrypted network traffic represented as 2D histograms of packet sizes and inter-arrival times. The data transforms raw network flows into image-like representations to facilitate application identification and VPN detection using computer vision techniques.
Use Cases
- Train a Convolutional Neural Network (CNN) to identify applications using the 2D histogram pixel intensity values
- Classify network traffic as VPN or non-VPN using the provided session-level labels
- Analyze network protocol behavior by comparing the visual clusters formed by packet size and timing features
Strengths
- 2D histograms with resolutions such as 32x32 or 64x64 pixels representing individual network flows
- 14 traffic labels covering categories like VoIP, File Transfer, and Instant Messaging across VPN and non-VPN sessions
- Features derived from packet size (0-1500 bytes) and inter-arrival time (0-15 seconds) distributions