Mapping the Invisible Internet: Framework and Dataset

This article presents a novel dataset focusing on the network layer of the Invisible Internet Project (I2P), where priorresearch has predominantly examined application layers like the darkweb. Data was collected through the SWARM-I2P framework deploying I2P routers as mapping agents, utilizing dynamic port mapping (30000-50000 range). Thedataset documents over 50,000 nodes, including 2,077 FastSet nodes and 2,331 high-capacity nodes characterized bybandwidth, latency (mean 121.21ms ±48.50), and uptime metrics. It contains 1,997 traffic records (1,003,032 pack-ets/bytes) and 4,222,793 records (2,147,585,625 packets/bytes), with geographic distributions for 3,444 peers showinging capacity metrics (mean 8.57 ±1.20). Collection methods included router console queries (127.0.0.1:port/tunnels),netDb analysis and passive monitoring with anonymized identifiers. Data is structured in CSV/TXT formats (Zen-odo) with collection scripts (GitHub). Potential applications include tunnel peer selection analysis, anonymity networkresilience studies and adversarial modeling.